U.S. patent application number 10/229339 was filed with the patent office on 2004-02-26 for apparatus, software, and methods for cardiac pulse detection using accelerometer data.
This patent application is currently assigned to Medtronic Physio-Control Manufacturing Corp.. Invention is credited to Hampton, David R., Jayne, Cynthia P., Joo, Tae H., Kelly, Patrick F., Lank, Paula, Nova, Richard C., O'Hearn, Patricia, Saltzstein, William E., Stickney, Ronald E..
Application Number | 20040039420 10/229339 |
Document ID | / |
Family ID | 31887655 |
Filed Date | 2004-02-26 |
United States Patent
Application |
20040039420 |
Kind Code |
A1 |
Jayne, Cynthia P. ; et
al. |
February 26, 2004 |
Apparatus, software, and methods for cardiac pulse detection using
accelerometer data
Abstract
A pulse detection apparatus, software, and method that uses
signal data obtained from an accelerometer placed on a patient's
body to detect the presence of a cardiac pulse. The accelerometer
is adapted to sense movement due to a cardiac pulse and produce
accelerometer signal data in response thereto. Processing circuitry
analyzes the accelerometer signal data for a feature indicative of
a cardiac pulse and determines whether a cardiac pulse is present
in the patient based on the feature. In one aspect, the feature may
be a temporal energy feature, such as a relative change in energy.
In another aspect, the feature may be a spectral energy feature
such as the energy or frequency of a peak in the energy spectrum of
the signal. In yet another aspect, the feature may be obtained by
comparing the accelerometer signal data with a
previously-identified pattern known to predict the presence of a
cardiac pulse. Multiple features may also be obtained and
classified to determine the presence of a cardiac pulse.
Inventors: |
Jayne, Cynthia P.; (Redmond,
WA) ; Stickney, Ronald E.; (Edmonds, WA) ;
Hampton, David R.; (Woodinville, WA) ; Lank,
Paula; (Renton, WA) ; O'Hearn, Patricia;
(Mercer Island, WA) ; Joo, Tae H.; (Redmond,
WA) ; Nova, Richard C.; (Kirkland, WA) ;
Kelly, Patrick F.; (Edmonds, WA) ; Saltzstein,
William E.; (Woodinville, WA) |
Correspondence
Address: |
Steven J. Shumaker
SHUMAKER & SIEFFERT, P.A.
8245 Seasons Parkwayt
Suite 105
St. Paul
MN
55125
US
|
Assignee: |
Medtronic Physio-Control
Manufacturing Corp.
|
Family ID: |
31887655 |
Appl. No.: |
10/229339 |
Filed: |
August 26, 2002 |
Current U.S.
Class: |
607/5 |
Current CPC
Class: |
A61B 2562/0219 20130101;
A61B 5/1107 20130101; A61B 5/7264 20130101; A61B 2562/0214
20130101; A61N 1/3925 20130101; A61B 5/6823 20130101 |
Class at
Publication: |
607/5 |
International
Class: |
A61N 001/39 |
Claims
The embodiments of the invention in which an exclusive property or
privilege is claimed are defined as follows:
1. A medical device for detecting the presence of a cardiac pulse,
comprising: (a) an accelerometer configured for placement on a
patient's body, the accelerometer being adapted to sense movement
in the patient's body due to a cardiac pulse and produce
accelerometer signal data in response thereto; and (b) processing
circuitry configured to analyze the accelerometer signal data for a
feature indicative of the presence of a cardiac pulse and determine
whether a cardiac pulse is present based on the feature.
2. The medical device of claim 1, in which the processing circuitry
is in communication with the accelerometer.
3. The medical device of claim 1, further comprising a display, in
which the processing circuitry is configured to automatically
report via the display whether a cardiac pulse is present in the
patient.
4. The medical device of claim 1, further comprising a display, in
which the processing circuitry is configured to automatically
prompt via the display the application of chest compressions or
cardiopulmonary resuscitation if the processing circuitry
determines that a cardiac pulse is not present in the patient.
5. The medical device of claim 1, further comprising a
defibrillation pulse generator in communication with the processing
circuitry for delivering a defibrillation pulse to the patient if
the processing circuitry determines that a cardiac pulse is not
present in the patient.
6. The medical device of claim 5, in which the medical device is an
automated external defibrillator.
7. The medical device of claim 6, in which the processing circuitry
is configured to automatically obtain and analyze the accelerometer
signal data to determine the presence of a cardiac pulse in the
patient.
8. The medical device of claim 5, further comprising an input
device that allows an operator of the medical device to initiate
delivery of the defibrillation pulse if the processing circuitry
determines that a cardiac pulse is not present in the patient.
9. The medical device of claim 1, in which the processing circuitry
is configured to determine the feature indicative of a cardiac
pulse from a temporal parameter in the accelerometer signal
data.
10. The medical device of claim 9, in which the feature indicative
of a cardiac pulse is an amplitude of the accelerometer signal
data, the processing circuitry being configured to compare the
amplitude to a threshold to determine whether a cardiac pulse is
present.
11. The medical device of claim 9, in which the feature indicative
of a cardiac pulse is an energy in the accelerometer signal data,
the processing circuitry being configured to compare the energy to
a threshold to determine whether a cardiac pulse is present.
12. The medical device of claim 9, in which the feature indicative
of a cardiac pulse is a derivative of the accelerometer signal
data, the processing circuitry being configured to compare the
derivative to a threshold to determine whether a cardiac pulse is
present.
13. The medical device of claim 9, in which the temporal parameter
is an energy in the accelerometer signal data, the processing
circuitry being configured to determine a relative change in energy
between an estimated first energy in the accelerometer signal data
and an estimated second energy in the accelerometer signal data,
and use the relative change in energy as the feature indicative of
a cardiac pulse.
14. The medical device of claim 13, in which the first energy is
estimated using a first set of accelerometer signal data and the
second energy is estimated using a second set of accelerometer
signal data, and in which the second set of accelerometer signal
data is obtained prior to the first set of accelerometer signal
data.
15. The medical device of claim 1, in which the processing
circuitry is configured to determine the feature indicative of a
cardiac pulse from a spectral parameter in the accelerometer signal
data.
16. The medical device of claim 15, in which the processing
circuitry is configured to calculate an energy spectrum of the
accelerometer signal data and locate a peak energy in the energy
spectrum, and in which the processing circuitry uses the energy
value of the located peak energy as the feature indicative of a
cardiac pulse.
17. The medical device of claim 15, in which the processing
circuitry is configured to calculate an energy spectrum of the
accelerometer signal data and locate a peak energy in the energy
spectrum, and in which the processing circuitry uses the frequency
at which the located peak energy occurs as the feature indicative
of a cardiac pulse.
18. The medical device of claim 1, in which the feature indicative
of the presence of a cardiac pulse is first feature, and in which
the processing circuitry is further configured to analyze the
accelerometer signal data for a second feature indicative of the
presence of a cardiac pulse, the processing circuitry being
configured to determine the presence of a cardiac pulse by
evaluating the first and second features.
19. The medical device of claim 18, in which the first feature and
the second feature are a temporal feature or a spectral feature
determined from the accelerometer signal data.
20. The medical device of claim 1, further comprising a display,
the processing circuitry being further configured to provide a
graph on the display showing a representation of the accelerometer
signal data.
21. The medical device of claim 1, further comprising an electrode
adapted to sense an electrocardiogram (ECG) signal in the patient
and communicate ECG signal data to the processing circuitry, the
processing circuitry being configured to analyze the ECG data in
connection with the accelerometer signal data to determine the
feature indicative of a cardiac pulse.
22. The medical device of claim 21, in which the processing
circuitry is further configured to determine the presence of a
ventricular complex in the ECG data and determine the presence of a
cardiac pulse in the patient if a ventricular complex occurs in the
ECG data within an expected time period in relation to a feature in
the accelerometer signal data that indicates a cardiac pulse.
23. The medical device of claim 21, in which the processing
circuitry is configured to analyze the ECG data and determine the
presence of a ventricular complex in the ECG data, the processing
circuitry being further configured to use the occurrence of a
ventricular complex to identify the accelerometer signal data to be
used in determining the presence of a cardiac pulse.
24. The medical device of claim 21, further comprising a display,
in which the processing circuitry is configured to prompt a message
via the display recommending application of chest compressions or
cardiopulmonary resuscitation to the patient if the processing
circuitry determines that a cardiac pulse is not present in the
patient and the ECG data obtained from the patient does not
indicate a cardiac rhythm appropriate for immediate treatment by
defibrillation therapy.
25. The medical device of claim 21, further comprising a
defibrillation pulse generator, in which the processing circuitry
is configured to instruct the defibrillation pulse generator to
generate a defibrillation pulse if the processing circuitry
determines that a cardiac pulse is not present in the patient and
that ECG data obtained from the patient indicates a cardiac rhythm
appropriate for treatment by defibrillation therapy.
26. The medical device of claim 25, further comprising a display,
in which the processing circuitry is configured to count the
delivery of defibrillation pulses to the patient and prompt a
message via the display recommending application of chest
compressions or cardiopulmonary resuscitation to the patient if the
number of defibrillation pulses delivered to the patient equals or
exceeds a predetermined number.
27. The medical device of claim 21, further comprising a display,
in which the processing circuitry is configured to prompt a message
via the display reporting whether the patient is in a state of
pulseless electrical activity (PEA).
28. The medical device of claim 27, in which the processing
circuitry determines the patient to be in a state of PEA if a
ventricular complex is found in the ECG data and a cardiac pulse is
not detected in the accelerometer signal data.
29. The medical device of claim 27, in which the processing
circuitry is further configured to analyze the patient's ECG data
for at least ventricular fibrillation (VF), ventricular tachycardia
(VT), and asystole, and if the patient is determined to be
pulseless and not in a VF, VT, or asystole condition, the
processing circuitry then prompting the message reporting that the
patient is in a state of PEA.
30. The medical device of claim 1, further comprising a display, in
which the processing circuitry is configured to prompt a message
via the display recommending application of rescue breathing
therapy to the patient if a cardiac pulse is not present and the
patient is not breathing.
31. The medical device of claim 1, in which the processing
circuitry is configured to analyze the accelerometer signal data
for a feature indicative of the presence of a cardiac pulse by
comparing the accelerometer signal data to a previously-identified
accelerometer signal data pattern known to predict the presence of
a cardiac pulse.
32. The medical device of claim 31, in which the comparison
produces a pattern match statistic that is the feature indicative
of the presence of a cardiac pulse, the processing circuitry being
further configured to compare the feature to a predetermined
pattern match threshold to determine whether a cardiac pulse is
present in the patient.
33. The medical device of claim 31, further comprising a display,
in which the processing circuitry is further configured to
automatically prompt a message via the display reporting whether a
cardiac pulse is present in the patient.
34. The medical device of claim 31, further comprising an electrode
adapted to sense an electrocardiogram (ECG) signal in the patient
and communicate ECG signal data to the processing circuitry, the
processing circuitry being configured to analyze the ECG data and
select accelerometer signal data corresponding in time with a
ventricular complex in the ECG data for the analysis of the
accelerometer signal data.
35. The medical device of claim 1, in which the processing
circuitry is further configured to report the return of spontaneous
circulation in the patient if a cardiac pulse is determined present
in the patient after delivery of defibrillation therapy to the
patient.
36. An electrotherapy device, comprising: (a) an accelerometer
configured for placement on a patient's body, the accelerometer
being adapted to sense movement in the patient's body due to a
cardiac pulse and produce accelerometer signal data in response
thereto; (b) an electrotherapy generator adapted for delivering
electrotherapy to the patient; and (c) processing circuitry
configured to analyze the accelerometer signal data for a feature
indicative of the presence of a cardiac pulse in the patient and
determine the presence of a cardiac pulse based on the feature, the
processing circuitry being further configured to prompt the
delivery of an electrotherapy to the patient based on the presence
of a cardiac pulse.
37. The electrotherapy device of claim 36, in which the processing
circuitry is in communication with the accelerometer and the
electrotherapy generator.
38. The electrotherapy device of claim 36, further comprising an
electrode adapted to sense an electrocardiogram (ECG) signal in the
patient and communicate ECG signal data to the processing
circuitry, the processing circuitry being further configured to
analyze the patient's ECG signal data for ventricular tachycardia
and prompt the delivery of defibrillation therapy to the patient if
the patient is determined to be pulseless and experiencing
ventricular tachycardia.
39. The electrotherapy device of claim 38, in which the processing
circuitry is configured to prompt the delivery of defibrillation
therapy if the patient is determined to be pulseless and
experiencing ventricular tachycardia with a rate exceeding 100
beats per minute.
40. The electrotherapy device of claim 36, further comprising an
electrode adapted to sense an electrocardiogram (ECG) signal in the
patient and communicate ECG signal data to the processing
circuitry, the processing circuitry being further configured to
analyze the patient's ECG signal data for at least ventricular
fibrillation (VF), ventricular tachycardia (VT), and asystole, and
if the patient is determined to be pulseless and not in a VF, VT,
or asystole condition, the processing circuitry then being
configured to prompt delivery of electrotherapy designed
specifically for pulseless electrical activity (PEA).
41. The electrotherapy device of claim 36, the processing circuitry
being further configured to report the return of spontaneous
circulation in the patient if a cardiac pulse is determined present
in the patient after delivery of electrotherapy to the patient.
42. The electrotherapy device of claim 36, further comprising an
electrode adapted to sense an electrocardiogram (ECG) signal in the
patient and communicate ECG signal data to the processing
circuitry, the processing circuitry being further configured to
analyze the patient's ECG signal data for one or more of
ventricular fibrillation (VF), ventricular tachycardia (VT),
asystole, and pulseless electrical activity (PEA), and prompt a
report of VF, VT, asystole, or PEA, if detected and if the patient
is determined to be pulseless.
43. The electrotherapy device of claim 42, in which the processing
circuitry determines the patient to be in a state of PEA if a
ventricular complex is found in the ECG signal data and the patient
is determined to be pulseless.
44. The electrotherapy device of claim 36, in which the
electrotherapy generator and the processing circuitry are
implemented in an automated external defibrillator.
45. The electrotherapy device of claim 44, further comprising a
display, in which the processing circuitry is configured to
automatically prompt via the display the delivery of chest
compressions or cardiopulmonary resuscitation to the patient if the
patient is determined to be pulseless.
46. An electrotherapy device, comprising: (a) an accelerometer
configured for placement on a patient's body, the accelerometer
being adapted to sense movement in the patient's body due to a
cardiac pulse and produce accelerometer signal data in response
thereto; (b) an electrotherapy generator for delivering pacing
stimuli to the patient; and (c) processing circuitry configured to
analyze the accelerometer signal data and determine whether a
cardiac pulse occurred in the patient following the delivery of a
pacing stimulus to the patient.
47. The electrotherapy device of claim 46, in which the processing
circuitry is configured to increase the current of further pacing
stimuli to be delivered to the patient if a cardiac pulse did not
occur in the patient following the delivery of the pacing
stimulus.
48. The electrotherapy device of claim 46, in which the
electrotherapy generator is configured to deliver pacing stimuli to
the patient two or more times and the processing circuitry is
configured to analyze the accelerometer signal data to determine
whether a cardiac pulse occurred after the delivery of each pacing
stimulus, the current of further pacing stimuli to be delivered to
the patient being increased if a cardiac pulse does not
consistently occur in the patient after the delivery of each pacing
stimulus.
49. The electrotherapy device of claim 48, in which prior to the
current of the pacing stimuli being increased, the processing
circuitry is configured to prompt a user of the device to increase
the pacing stimuli current.
50. An article comprising a storage medium having device-executable
instructions stored thereon, in which when the instructions are
executed by at least one device, they result in: (a) obtaining
accelerometer signal data from an accelerometer placed on a
patient's body; (b) analyzing the accelerometer signal data for a
feature indicative of the presence of a cardiac pulse; and (c)
determining whether a cardiac pulse is present in the patient based
on the feature in the accelerometer signal data.
51. The article of claim 50, in which analyzing the accelerometer
signal data includes evaluating a temporal parameter in the
accelerometer signal data.
52. The article of claim 51, in which evaluating a temporal
parameter in the accelerometer signal data includes: (a) estimating
an instantaneous energy in the accelerometer signal data; (b)
estimating a background energy in the accelerometer signal data;
and (c) comparing the instantaneous energy with the background
energy to produce the feature indicative of the presence of a
cardiac pulse.
53. The article of claim 50, in which analyzing the accelerometer
signal data includes evaluating a spectral parameter in the
accelerometer signal data.
54. The article of claim 53, in which evaluating a spectral
parameter in the accelerometer signal data includes calculating an
energy spectrum of the accelerometer signal data and evaluating the
energy spectrum to locate a peak energy value, the instructions
when executed further resulting in using the located peak energy
value as the feature indicative of the presence of a cardiac pulse
and determining whether a cardiac pulse is present in the patient
by comparing the located peak energy value with a threshold energy
value.
55. The article of claim 53, in which evaluating a spectral
parameter in the accelerometer signal data includes calculating an
energy spectrum of the accelerometer signal data, evaluating the
energy spectrum to locate a peak energy value, and determining the
frequency at which the peak energy value occurs, the instructions
when executed further resulting in using the frequency of the peak
energy value as the feature indicative of the presence of a cardiac
pulse and determining whether a cardiac pulse is present in the
patient by comparing the frequency of the peak energy value with a
threshold frequency.
56. The article of claim 50, in which executing the instructions
further results in: (a) repeating the steps of obtaining
accelerometer signal data, analyzing the accelerometer signal data
for a feature, and determining whether a cardiac pulse is present
based on the feature, to produce two or more preliminary
determinations of the presence of a cardiac pulse; and (b)
determining whether a cardiac pulse is present in the patient based
on the number of preliminary determinations indicating the presence
of a cardiac pulse.
57. The article of claim 50, in which analyzing the accelerometer
signal data includes comparing the accelerometer signal data to a
previously-identified accelerometer signal data pattern known to
predict the presence of a cardiac pulse.
58. The article of claim 57, in which the comparison produces a
pattern match statistic that is the feature indicative of the
presence of a cardiac pulse, the instructions when executed further
resulting in comparing the feature to a predetermined pattern match
threshold to determine whether a cardiac pulse is present in the
patient.
59. The article of claim 57, in which executing the instructions
further results in analyzing the accelerometer signal data for two
or more features indicative of the presence of a cardiac pulse, in
which one of the features is determined from the comparison of the
accelerometer signal data with a previously-identified
accelerometer signal data pattern and in which one of the other
features is determined from an evaluation of an amplitude of the
accelerometer signal data or an energy in the accelerometer signal
data.
60. The article of claim 57, in which executing the instructions
further results in obtaining electrocardiogram (ECG) data from the
patient, and in which analyzing the obtained accelerometer signal
data for a feature indicative of the presence of a cardiac pulse
further includes determining whether a ventricular complex occurred
in the ECG data.
61. The article of claim 60, in which executing the instructions
further results in locating a ventricular complex in the ECG data
and selecting accelerometer signal data for the pattern match
comparison based on the location of the ventricular complex.
62. The article of claim 60, in which executing the instructions
further results in determining whether the patient is in a state of
pulseless electrical activity (PEA).
63. The article of claim 62, in which the patient is determined to
be in a state of PEA if a ventricular complex is found in the ECG
data and the patient is determined to be pulseless.
64. The article of claim 62, in which executing the instructions
further results in analyzing the patient's ECG data for at least
ventricular fibrillation (VF), ventricular tachycardia (VT), and
asystole, and determining that the patient is in a state of PEA if
the patient is determined to be pulseless and not in a VF, VT, or
asystole condition.
65. An article comprising a storage medium having device-executable
instructions stored thereon, in which when the instructions are
executed by at least one device, they result in: (a) obtaining
accelerometer signal data from an accelerometer placed on a
patient's body; (b) estimating a first energy in the accelerometer
signal data; (c) estimating a second energy in the accelerometer
signal data; (d) determining a relative change in energy between
the first energy and the second energy; and (e) determining the
presence of a cardiac pulse in the patient based on the determined
relative change in energy.
66. The article of claim 65, in which the first energy is estimated
using a first set of accelerometer signal data and the second
energy is estimated using a second set of accelerometer signal
data, and in which the second set of accelerometer signal data is
obtained prior to the first set of accelerometer signal data.
67. The article of claim 65, in which executing the instructions
further results in: (a) calculating an energy spectrum of the
accelerometer signal data; (b) evaluating the energy spectrum for a
spectral energy feature indicative of the presence of a cardiac
pulse; and (c) determining the presence of a cardiac pulse in the
patient based on the determined relative change in energy and the
spectral energy feature.
68. An article comprising a storage medium having device-executable
instructions stored thereon, in which when the instructions are
executed by at least one device, they result in: (a) obtaining
accelerometer signal data from an accelerometer placed on a
patient's body; (b) calculating an energy spectrum of the
accelerometer signal data; (c) evaluating the energy spectrum for a
spectral energy feature indicative of the presence of a cardiac
pulse; and (d) determining the presence of a cardiac pulse in the
patient based on the spectral energy feature.
69. The article of claim 68, in which the spectral energy feature
is a peak energy value in the energy spectrum.
70. The article of claim 69, in which determining the presence of a
cardiac pulse includes comparing the peak energy value with a
threshold energy value.
71. The article of claim 69, in which determining the presence of a
cardiac pulse includes evaluating the frequency at which the peak
energy value occurs in the energy spectrum.
72. The article of claim 71, in which evaluating the frequency at
which the peak energy value occurs includes comparing the frequency
of the peak energy value with a threshold frequency.
73. The article of claim 68, in which executing the instructions
further results in identifying a set of accelerometer signal data
that has a higher likelihood of indicating the presence of a
cardiac pulse, and using the set of accelerometer signal data to
calculate the energy spectrum.
74. The article of claim 68, in which the spectral energy feature
is a first spectral energy feature, the instructions when executed
further resulting in evaluating the energy spectrum for a second
spectral energy feature indicative of the presence of a cardiac
pulse, in which determining the presence of a cardiac pulse in the
patient is based on the first and second spectral energy
features.
75. The article of claim 74, in which the first spectral energy
feature is a peak energy value in the energy spectrum, and in which
the second spectral energy feature is the frequency at which a peak
energy value occurs in the energy spectrum.
76. The article of claim 75, in which determining the presence of a
cardiac pulse in the patient includes comparing the first spectral
energy feature with a threshold energy value, and comparing the
second spectral energy feature with a threshold frequency.
77. The article of claim 68, in which executing the instructions
further results in evaluating a temporal parameter in the
accelerometer signal data for a temporal feature, in which
determining the presence of a cardiac pulse in the patient is based
on the spectral energy feature and the temporal feature.
78. The article of claim 77, in which the temporal parameter is
energy and the temporal energy feature in determined by estimating
a first energy in the accelerometer signal data, estimating a
second energy in the accelerometer signal data, and determining a
relative change in energy between the first energy and the second
energy.
79. The article of claim 78, in which the first energy is estimated
using a first set of accelerometer signal data and the second
energy is estimated using a second set of accelerometer signal
data, and in which the second set of accelerometer signal data is
obtained prior to the first set of accelerometer signal data.
80. The article of claim 77, in which the temporal feature is based
on an estimated energy in the accelerometer signal data, and in
which the spectral energy feature is based on a peak energy value
in the energy spectrum.
81. The article of claim 77, in which the temporal feature and
spectral energy feature are jointly classified in a
multi-dimensional classifier to determine whether a cardiac pulse
is present in the patient.
82. An article comprising a storage medium having device-executable
instructions stored thereon, in which when the instructions are
executed by at least one device, they result in: (a) delivering a
pacing stimulus to the patient; (b) obtaining accelerometer signal
data from an accelerometer placed on the patient's body; (c)
analyzing the accelerometer signal data to determine whether a
cardiac pulse occurred in the patient after delivery of the pacing
stimulus; and (d) if a cardiac pulse did not occur in the patient
after delivery of the pacing stimulus, increasing the current of
further pacing stimuli to be delivered to the patient.
83. The article of claim 82, in which executing the instructions
further results in repeating steps (a)-(d) until a cardiac pulse
occurs after delivery of the pacing stimulus.
84. The article of claim 82, in which executing the instructions
results in delivering pacing stimuli to the patient two or more
times and analyzing the accelerometer signal data to determine
whether a cardiac pulse occurred after the delivery of each pacing
stimulus, the current of further pacing stimuli to be delivered to
the patient being increased if a cardiac pulse does not
consistently occur in the patient after the delivery of each pacing
stimulus.
85. The article of claim 84, in which prior to the current of the
pacing stimuli being increased, executing the instructions results
in prompting a user of the device to increase the pacing stimuli
current.
86. The article of claim 84, in which executing the instructions
further results in repeating the delivery of pacing stimuli and
increasing the current of the pacing stimuli until a cardiac pulse
consistently occurs in the patient after the delivery of each
pacing stimulus.
87. A method of determining the presence of a cardiac pulse,
comprising: (a) obtaining accelerometer signal data from an
accelerometer placed on a patient's body; (b) analyzing the
accelerometer signal data for a feature indicative of the presence
of a cardiac pulse; and (c) determining whether a cardiac pulse is
present in the patient based on the feature in the accelerometer
signal data.
88. The method of claim 87, in which analyzing the accelerometer
signal data includes evaluating a temporal parameter in the
accelerometer signal data.
89. The method of claim 88, in which evaluating a temporal
parameter in the accelerometer signal data includes: (a) estimating
an instantaneous energy in the accelerometer signal data; (b)
estimating a background energy in the accelerometer signal data;
and (c) comparing the instantaneous energy with the background
energy to produce the feature indicative of the presence of a
cardiac pulse.
90. The method of claim 87, in which analyzing the accelerometer
signal data includes evaluating a spectral parameter in the
accelerometer signal data.
91. The method of claim 90, in which evaluating a spectral
parameter in the accelerometer signal data includes calculating an
energy spectrum of the accelerometer signal data and evaluating the
energy spectrum to locate a peak energy value, in which the located
peak energy value is used as the feature indicative of the presence
of a cardiac pulse, and in which determining whether a cardiac
pulse is present in the patient includes comparing the located peak
energy value with a threshold energy value.
92. The method of claim 90, in which evaluating a spectral
parameter in the accelerometer signal data includes calculating an
energy spectrum of the accelerometer signal data, evaluating the
energy spectrum to locate a peak energy value, and determining the
frequency at which the peak energy value occurs, in which the
frequency of the peak energy value is used as the feature
indicative of the presence of a cardiac pulse, and in which
determining whether a cardiac pulse is present in the patient
includes comparing the frequency of the peak energy value with a
threshold frequency.
93. The method of claim 87, further comprising: (a) repeating the
steps of obtaining accelerometer signal data, analyzing the
accelerometer signal data for a feature, and determining whether a
cardiac pulse is present based on the feature, to produce two or
more preliminary determinations of the presence of a cardiac pulse;
and (b) determining whether a cardiac pulse is present in the
patient based on the number of preliminary determinations
indicating the presence of a cardiac pulse.
94. The method of claim 87, in which analyzing the accelerometer
signal data includes comparing the accelerometer signal data to a
previously-identified accelerometer signal data pattern known to
predict the presence of a cardiac pulse.
95. The method of claim 94, in which the comparison produces a
pattern match statistic that is the feature indicative of the
presence of a cardiac pulse, the method further comprising
comparing the feature to a predetermined pattern match threshold to
determine whether a cardiac pulse is present in the patient.
96. The method of claim 94, further comprising analyzing the
accelerometer signal data for two or more features indicative of
the presence of a cardiac pulse, in which one of the features is
determined from the comparison of the accelerometer signal data
with a previously-identified accelerometer signal data pattern and
in which one of the other features is determined from an evaluation
of an amplitude of the accelerometer signal data or an energy in
the accelerometer signal data.
97. The method of claim 94, further comprising obtaining
electrocardiogram (ECG) data from the patient, in which analyzing
the obtained accelerometer signal data for a feature indicative of
the presence of a cardiac pulse further includes determining
whether a ventricular complex occurred in the ECG data.
98. The method of claim 97, further comprising locating a
ventricular complex in the ECG data and selecting accelerometer
signal data for the pattern match comparison based on the location
of the ventricular complex.
99. The method of claim 97, further comprising determining whether
the patient is in a state of pulseless electrical activity.
100. The method of claim 99, in which the patient is determined to
be in a state of PEA if a ventricular complex is found in the ECG
data and the patient is determined to be pulseless.
101. The method of claim 99, further comprising analyzing the
patient's ECG data for at least ventricular fibrillation (VF),
ventricular tachycardia (VT), and asystole, and determining that
the patient is in a state of PEA if the patient is determined to be
pulseless and not in a VF, VT, or asystole condition.
102. A method of determining the presence of a cardiac pulse,
comprising: (a) obtaining accelerometer signal data from an
accelerometer placed on a patient's body; (b) estimating a first
energy in the accelerometer signal data; (c) estimating a second
energy in the accelerometer signal data; (d) determining a relative
change in energy between the first energy and the second energy;
and (e) determining the presence of a cardiac pulse in the patient
based on the determined relative change in energy.
103. The method of claim 102, in which the first energy is
estimated using a first set of accelerometer signal data and the
second energy is estimated using a second set of accelerometer
signal data, and in which the second set of accelerometer signal
data is obtained prior to the first set of accelerometer signal
data.
104. The method of claim 102, further comprising: (a) calculating
an energy spectrum of the accelerometer signal data; (b) evaluating
the energy spectrum for a spectral energy feature indicative of the
presence of a cardiac pulse; and (c) determining the presence of a
cardiac pulse in the patient based on the determined relative
change in energy and the spectral energy feature.
105. A method of determining the presence of a cardiac pulse,
comprising: (a) obtaining accelerometer signal data from an
accelerometer placed on a patient's body; (b) calculating an energy
spectrum of the accelerometer signal data; (c) evaluating the
energy spectrum for a spectral energy feature indicative of the
presence of a cardiac pulse; and (d) determining the presence of a
cardiac pulse in the patient based on the spectral energy
feature.
106. The method of claim 105, in which the spectral energy feature
is a peak energy value in the energy spectrum.
107. The method of claim 106, in which determining the presence of
a cardiac pulse includes comparing the peak energy value with a
threshold energy value.
108. The method of claim 106, in which determining the presence of
a cardiac pulse includes evaluating the frequency at which the peak
energy value occurs in the energy spectrum.
109. The method of claim 108, in which evaluating the frequency at
which the peak energy value occurs includes comparing the frequency
of the peak energy value with a threshold frequency.
110. The method of claim 105, further comprising identifying a set
of accelerometer signal data that has a higher likelihood of
indicating the presence of a cardiac pulse, and using the set of
accelerometer signal data to calculate the energy spectrum.
111. The method of claim 105, in which the spectral energy feature
is a first spectral energy feature, the method further comprising
evaluating the energy spectrum for a second spectral energy feature
indicative of the presence of a cardiac pulse, in which determining
the presence of a cardiac pulse in the patient is based on the
first and second spectral energy features.
112. The method of claim 111, in which the first spectral energy
feature is a peak energy value in the energy spectrum, and in which
the second spectral energy feature is the frequency at which a peak
energy value occurs in the energy spectrum.
113. The method of claim 112, in which determining the presence of
a cardiac pulse in the patient includes comparing the first
spectral energy feature with a threshold energy value, and
comparing the second spectral energy feature with a threshold
frequency.
114. The method of claim 105, further comprising evaluating a
temporal parameter in the accelerometer signal data for a temporal
feature, in which determining the presence of a cardiac pulse in
the patient is based on the spectral energy feature and the
temporal feature.
115. The method of claim 114, in which the temporal parameter is
energy and the temporal energy feature in determined by estimating
a first energy in the accelerometer signal data, estimating a
second energy in the accelerometer signal data, and determining a
relative change in energy between the first energy and the second
energy.
116. The method of claim 115, in which the first energy is
estimated using a first set of accelerometer signal data and the
second energy is estimated using a second set of accelerometer
signal data, and in which the second set of accelerometer signal
data is obtained prior to the first set of accelerometer signal
data.
117. The method of claim 114, in which the temporal feature is
based on an estimated energy in the accelerometer signal data, and
in which the spectral energy feature is based on a peak energy
value in the energy spectrum.
118. The method of claim 114, in which the temporal feature and
spectral energy feature are jointly classified in a
multi-dimensional classifier to determine whether a cardiac pulse
is present in the patient.
119. A method for delivering electrotherapy that provides pacing
stimuli and seeks capture of a cardiac pulse in a patient, the
method comprising: (a) delivering a pacing stimulus to the patient;
(b) obtaining accelerometer signal data from an accelerometer
placed on the patient's body; (c) analyzing the accelerometer
signal data to determine whether a cardiac pulse occurred in the
patient after delivery of the pacing stimulus; and (d) if a cardiac
pulse did not occur in the patient after delivery of the pacing
stimulus, increasing the current of further pacing stimuli to be
delivered to the patient.
120. The method of claim 119, further comprising repeating steps
(a)-(d) until a cardiac pulse occurs after delivery of the pacing
stimulus.
121. The method of claim 119, in which pacing stimuli is delivered
to the patient two or more times and the accelerometer signal data
is analyzed to determine whether a cardiac pulse occurred after the
delivery of each pacing stimulus, and in which the current of
further pacing stimuli to be delivered to the patient is increased
if a cardiac pulse does not consistently occur in the patient after
the delivery of each pacing stimulus.
122. The method of claim 121, further comprising prompting a user
of the device to increase the pacing stimuli current prior to the
current of the pacing stimuli being increased.
123. The method of claim 121, further comprising repeating the
delivery of pacing stimuli and increasing the current of the pacing
stimuli until a cardiac pulse consistently occurs in the patient
after the delivery of each pacing stimulus.
Description
FIELD OF THE INVENTION
[0001] The present invention relates to detection of cardiac
activity in a patient, and more specifically, to detection of a
cardiac pulse and use of pulse detection in delivering therapy.
BACKGROUND OF THE INVENTION
[0002] The presence of a cardiac pulse in a patient is typically
detected by palpating the patient's neck and sensing changes in the
volume of the patient's carotid artery due to blood pumped from the
patient's heart. When the heart's ventricles contract during a
heartbeat, a pressure wave is sent throughout the patient's
peripheral circulation system. A carotid pulse waveform rises with
the ventricular ejection of blood at systole and peaks when the
pressure wave from the heart reaches a maximum. The carotid pulse
falls off again as the pressure subsides toward the end of the
pulse.
[0003] The absence of a detectable cardiac pulse in a patient is a
strong indicator of cardiac arrest. Cardiac arrest is a
life-threatening medical condition in which the patient's heart
fails to provide sufficient blood flow to support life. During
cardiac arrest, the electrical activity of the heart may be
disorganized (ventricular fibrillation), too rapid (ventricular
tachycardia), absent (asystole), or organized at a normal or slow
heart rate without producing sufficient blood flow (pulseless
electrical activity).
[0004] The form of therapy to be provided to a patient in cardiac
arrest depends, in part, on an assessment of the patient's cardiac
condition. For example, a caregiver may apply a defibrillation
shock to a patient experiencing ventricular fibrillation (VF) or
ventricular tachycardia (VT) to stop the unsynchronized or rapid
electrical activity and allow a perfusing rhythm to return.
External defibrillation, in particular, is provided by applying a
strong electric pulse to the patient's heart through electrodes
placed on the surface of the patient's body. If the patient lacks a
detectable pulse and is experiencing asystole or pulseless
electrical activity (PEA), a caregiver may perform cardiopulmonary
resuscitation (CPR), which causes some blood to flow in the
patient.
[0005] Before providing therapy such as defibrillation or CPR to a
patient, a caregiver must first confirm that the patient is in
cardiac arrest. In general, external defibrillation is suitable
only for patients that are unconscious, apneic, pulseless, and in
VF or VT. Medical guidelines indicate that the presence or absence
of a cardiac pulse in a patient should be determined within 10
seconds. See "American Heart Guidelines 2000 For Cardiopulmonary
Resuscitation and Emergency Cardiovascular Care, Part 3: Adult
Basic Life Support," Circulation 102 Suppl. I:-22 to I-59,
2000.
[0006] Unfortunately, under the pressure and stress of an emergency
situation, it can be extremely difficult for first-responding
caregivers with little or no medical training to consistently and
accurately detect a cardiac pulse in a patient (e.g., by palpating
the carotid artery) in a short amount of time such as 10 seconds.
See Eberle B. et al. "Checking the Carotid Pulse Diagnostic
Accuracy of First Responders in Patients With and Without a Pulse,"
Resuscitation 33:107-116, 1996. Nevertheless, because time is of
the essence in treating cardiac arrest, a caregiver may rush the
preliminary evaluation, incorrectly conclude that the patient has
no pulse, and proceed to provide therapy, such as defibrillation,
when in fact the patient has a pulse. In other circumstances, the
caregiver may incorrectly conclude that the patient has a pulse and
erroneously withhold defibrillation therapy. A need therefore
exists for a method and apparatus that quickly, accurately, and
automatically determines whether a cardiac pulse is present in a
patient, particularly to prompt a caregiver to provide appropriate
therapy in an emergency situation.
SUMMARY OF THE INVENTION
[0007] The present invention provides pulse detection apparatus,
software, and methods that use signal data obtained from an
accelerometer placed on a patient's body. The accelerometer is
adapted to sense movement in the patient's body due to a cardiac
pulse and produce accelerometer signal data in response thereto.
Processing circuitry is configured to analyze the accelerometer
signal data for a feature indicative of the presence of a cardiac
pulse. The processing circuitry then determines whether a cardiac
pulse is present in the patient based on the feature.
[0008] A device constructed according to the invention may further
comprise a display that automatically reports whether a cardiac
pulse is present in the patient. The device may also include a
defibrillation pulse generator that delivers a defibrillation pulse
to the patient if the processing circuitry in the device determines
that a cardiac pulse is not present in the patient.
[0009] In one aspect, the feature indicative of a cardiac pulse may
be a temporal parameter. For example, the processing circuitry may
determine a relative change in energy between a first energy in the
accelerometer signal data and a second energy in the accelerometer
signal data, the relative change in energy constituting the feature
indicative of a cardiac pulse. In that regard, the first and second
energy may be estimated using segments of accelerometer signal data
that are obtained at different times.
[0010] In another aspect, the feature indicative of a cardiac pulse
may be a spectral parameter. In one exemplary implementation, the
processing circuitry calculates an energy spectrum of the
accelerometer signal data and locates a peak energy in the energy
spectrum. The energy value of the located peak is used as the
feature indicative of a cardiac pulse. In another implementation,
the frequency of a located peak energy is used as the feature
indicative of a cardiac pulse. In either case, a cardiac pulse may
be determined by comparing the feature with a predetermined
threshold. Multiple features may also be obtained from the
accelerometer signal data and classified to determine the presence
of a cardiac pulse.
[0011] In yet another aspect, electrocardiogram (ECG) signals may
be used in the analysis of the accelerometer signal data. A device
constructed according to one implementation of the invention may
determine whether a ventricular complex, such as a QRS complex, is
present in the ECG data, and if so, select and analyze a segment of
accelerometer signal data corresponding in time to the detected
ventricular complex. In another implementation, the presence of a
ventricular complex may be used to verify the detection of a
cardiac pulse by determining whether a ventricular complex occurred
in the ECG data within an expected time period in relation to the
feature in the accelerometer signal data. An ECG analysis may also
be used to determine whether defibrillation pulse therapy is
appropriate for a patient that is determined to be pulseless. In
other applications, the device may recommend providing chest
compressions or cardiopulmonary resuscitation (CPR) to the
patient.
[0012] In yet another aspect, the feature indicative of the
presence of a cardiac pulse may be obtained by comparing the
accelerometer signal data with a previously-identified
accelerometer signal data pattern known to predict the presence of
a cardiac pulse. The comparison may produce a pattern match
statistic that is compared with a predetermined pattern match
threshold to determine whether a cardiac pulse is present.
[0013] In yet further implementations, ECG data obtained from the
patient with the accelerometer signal data may be used to assess
the patient's cardiac activity. If, for instance, ventricular
tachycardia is detected and the patient is determined to be
pulseless, the device may prompt the delivery of defibrillation
therapy to the patient. The device may be further configured to
determine whether the patient is experiencing ventricular
fibrillation, ventricular tachycardia, or asystole, and if the
patient is not in a VF, VT, or asystole condition and is pulseless,
the device may prompt delivery of electrotherapy designed
specifically for pulseless electrical activity (PEA).
[0014] Embodiments of the invention intended for trained medical
personnel may also provide a graph of the accelerometer signal data
that is representative of the presence or absence of a pulse in the
patient. For example, the accelerometer signal data may be shown as
a waveform on a computer screen. The accelerometer signal data may
also be displayed as a bar whose length fluctuates according to the
accelerometer signal data. Other known display formats may also be
used.
BRIEF DESCRIPTION OF THE DRAWINGS
[0015] The foregoing aspects and many of the attendant advantages
of this invention will become more readily appreciated as the same
become better understood by reference to the following detailed
description, when taken in conjunction with the accompanying
drawings, wherein:
[0016] FIG. 1 is a graph depicting an electrocardiogram (ECG)
waveform for three consecutive heartbeats of a human patient;
[0017] FIG. 2 is a graph depicting an accelerometer signal waveform
for three consecutive heartbeats of a human patient, in which the
signal is obtained from an accelerometer placed on the patient's
body;
[0018] FIG. 3 is a pictorial diagram of a defibrillator,
electrodes, and accelerometer constructed in accordance with one
embodiment of the present invention and attached to a patient;
[0019] FIG. 4 is a simplified side cross-sectional view of a sensor
in one embodiment of the accelerometer shown in FIG. 3;
[0020] FIG. 5 is a block diagram of major components of a
defibrillator as shown in FIG. 3;
[0021] FIG. 6 is a flow diagram of a pulse detection process
performed by a defibrillator as shown in FIG. 3, in which an
analysis of temporal energy in accelerometer signal data obtained
from a patient is performed;
[0022] FIG. 7 is a flow diagram of another pulse detection process
performed by a defibrillator as shown in FIG. 3, in which a
spectral peak frequency analysis of accelerometer signal data is
performed;
[0023] FIG. 8 is a flow diagram of another pulse detection process
performed by a defibrillator as shown in FIG. 3, in which a
spectral peak energy analysis of accelerometer signal data is
performed;
[0024] FIG. 9 is a flow diagram of yet another pulse detection
process performed by a defibrillator as shown in FIG. 3 that
incorporates aspects of the pulse detection processes shown in
FIGS. 6, 7 and 8;
[0025] FIG. 10 is a flow diagram of a pulse detection process
performed by a defibrillator as shown in FIG. 3 that includes
analysis of one or more segments of accelerometer signal data;
[0026] FIG. 11 is a flow diagram of a pulse rate analysis performed
with the pulse detection process shown in FIG. 10;
[0027] FIG. 12 is a flow diagram of another pulse detection process
performed in accordance with the present invention in which an
accelerometer signal pattern analysis is performed;
[0028] FIG. 13 is a flow diagram of a procedure implemented by a
defibrillator as shown in FIG. 3 that incorporates a pulse
detection process provided by the present invention;
[0029] FIG. 14 is a flow diagram of another procedure implemented
by a defibrillator as shown in FIG. 3 that incorporates a pulse
detection process provided by the present invention;
[0030] FIG. 15 is a flow diagram of still another procedure
implemented by a defibrillator as shown in FIG. 3 that incorporates
a pulse detection process provided by the present invention;
[0031] FIG. 16 is a flow diagram of an auto-capture detection
process for cardiac pacing that uses a pulse detection process of
the present invention; and
[0032] FIG. 17 is a flow diagram of a patient condition advisory
process for use in a medical device that incorporates a pulse
detection process of the present invention.
DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENT
[0033] An electrocardiogram (ECG) waveform, as shown in FIG. 1,
depicts the electrical activity of a patient's heart. A patient
experiencing normal cardiac activity will exhibit an ECG waveform
having standard identifiable features. The portion of the ECG
waveform representing depolarization of the atrial muscle fibers is
referred to as the "P" wave, as shown in FIG. 1. Depolarization of
the ventricular muscle fibers is collectively represented by the
"Q," "R," and "S" waves. Finally, the portion of the waveform
representing repolarization of the ventricular muscle fibers is
known as the "T" wave. Between heartbeats, a normal ECG waveform
generally returns to an isopotential level.
[0034] The contraction and release of cardiac muscle in normal
cardiac activity produces vibrations through the chest cavity that
can be detected on the surface of the patient's body. Higher
frequency vibrations from the opening and closing of the patient's
heart valves are also detectable by equipment placed on the
patient's body. Conventionally, a physician listens to a patient's
heartbeat by placing a stethoscope on the patient's chest. A
transducer in the stethoscope senses the sound produced by the
heartbeat and delivers an acoustic signal that the physician can
hear. Less technological but sometimes effective is simply to place
a hand on the patient's chest. Although this does not substitute
for checking the patient's pulse by palpating an appropriate
pressure point (e.g., the carotid artery), vibrations in the chest
wall may be detected.
[0035] The present invention is directed to a method and apparatus
for cardiac pulse detection using a signal generated by an
accelerometer placed on the patient's chest. When the accelerometer
is placed on the patient's body, vibrations in the chest wall
caused by the patient's heart cause the accelerometer to output an
electric signal. This electric signal is transmitted to processing
circuitry that analyzes the signal to determine whether a cardiac
pulse is present in the patient.
[0036] FIG. 2 depicts a waveform of accelerometer signal data
obtained from an accelerometer placed on the chest of a patient.
The timing of the accelerometer signal data depicted in FIG. 2
correlates with the timing of the ECG data shown in FIG. 1. It is
significant to note that the peak values in the accelerometer
signal data consistently occur in time relation to the QRS
complexes depicted in the ECG data. It is thus evident that the
accelerometer signal data includes features, much as ECG data, that
are indicative of the presence of a cardiac pulse in the
patient.
[0037] Although the present invention may be implemented in a
variety of applications, it is particularly suited for use in a
defibrillator, such as the defibrillator 10 shown in FIG. 3. In
FIG. 3, the defibrillator 10 is shown connected to a patient 18 via
defibrillation electrodes 12 and 14 placed on the skin of the
patient 18. The defibrillator 10 uses the defibrillation electrodes
12 and 14 to deliver defibrillation pulses to the patient 18. The
defibrillator 10 may also use the electrodes 12 and 14 to obtain
ECG signals from the patient 18.
[0038] FIG. 3 further illustrates an accelerometer 16 placed on the
patient 18. The accelerometer 16 is located on a flap connected to
the electrode 14 and is configured to detect cardiac vibrations in
the chest wall of the patient. Vibrations sensed by the sensor 16
are converted by the defibrillator 10 into digital accelerometer
signal data for processing. Alternatively, the accelerometer 16 may
be integrated with or attached to either or both of the electrodes
12 and 14, e.g., as shown by reference numeral 17. The
accelerometer 16 may also be separately attached to the patient 18
by one or more wires (not shown).
[0039] It is well understood that acceleration is a change in
velocity per unit time. An accelerometer is a sensor designed to
measure accelerations that result from forces applied to the
accelerometer. In particular, an accelerometer responds to the
component of acceleration corresponding to the accelerometer's
sensitive axis or axes.
[0040] The unit of acceleration equal to the average force of
gravity occurring at the Earth's surface is generally represented
by the letter g. A g is approximately equal to 9.8 m/s.sup.2.
Accelerometers are generally configured to output a voltage signal
that changes per g unit of acceleration sensed by the device
(generally specified in terms of mV/g).
[0041] One exemplary accelerometer that may be used in the present
invention is manufactured by Analog Devices of Norwood, Mass.,
under part number ADXL150, which is a low noise, low power, single
axis accelerometer. Analog Devices also manufactures a dual axis
accelerometer under part number ADXL250 which may also be used in
the present invention. Both devices have their sensitive axes in
the same plane as the chip on which it is made. Other commercially
available accelerometers may also be used, including 3-axis
accelerometers.
[0042] Accelerometers, such as the ADXL150 noted above, may be
fabricated using standard integrated circuit manufacturing methods.
The signal processing circuitry may be combined on the same chip
with the acceleration sensor. One suitable manufacturing method for
an acceleration sensor is depositing polysilicon on a sacrificial
oxide layer that is then etched away leaving the suspended sensor
element. FIG. 4 depicts a simplified view of one such sensor
structure. Actual accelerometers generally include multiple unit
cells for sensing acceleration.
[0043] The sensor depicted in FIG. 4 is a differential capacitor
sensor. The sensor includes a beam structure 20 that is anchored
via anchor points 21. Included in or attached to the beam structure
20 is a plate 22. The beam 20 and plate 22 move between the anchor
points 21 in response to acceleration.
[0044] The moving plate 22 is disposed between fixed plates 24 that
are anchored via anchor points 25. Movement of the plate 22 between
the fixed plates 24 changes the differential capacitance in the
sensor which is measured by signal processing circuitry.
[0045] The sensed changes in capacitance are converted to a voltage
that is output by the accelerometer. The ADXL150 accelerometer
noted above outputs a voltage V.sub.out that is a function of both
the acceleration input "a," a power supply voltage "V.sub.s," and a
"sensitivity" parameter as follows: 1 V out = Vs 2 - ( Sensitivity
.times. Vs 5 V .times. a )
[0046] The scale factor of an accelerometer specifies the voltage
change of the output per g of applied acceleration. In applications
for the present invention, the amount of acceleration due to
movements caused by cardiac pulses in the patient may be small.
Accordingly, it is preferable to adjust the scale factor of the
accelerometer to appropriately measure the lower g accelerations
due to a cardiac pulse. The output scale factor may be increased by
either programmable pins on the accelerometer itself or by using a
buffer amplifier external to the accelerometer.
[0047] Furthermore, in the present invention, a DC (gravity)
response from the accelerometer is not required as the movement of
interest is the vibration of the patient's chest wall due to
cardiac pulses. Accordingly, AC coupling can be used to connect the
accelerometer's output to an external amplifier. The use of AC
coupling virtually eliminates any zero g drift and maximizes the
gain of the external amplifier without signal clipping.
[0048] As will be discussed in greater detail below in regard to
FIG. 5, low-pass or band pass filtering of the output signal may be
used to reduce the measurement bandwidth, and hence reduce noise in
the signal. An improved signal-to-noise ratio in the signal can be
important when measuring low g accelerations. The signal-to-noise
ratio may also be improved by performing multiple measurements and
then computing an average signal level.
[0049] Persons having ordinary skill in the art will recognize that
the layout of the accelerometer sensor depicted in FIG. 4 is an
exemplary design only. The dimensions, shape, and construction of
the sensor may be modified according to known techniques as
required. Additional information regarding material and techniques
for constructing and using an accelerometer, including the AC
coupling discussed above, is available from Analog Devices. See,
e.g., the technical specification sheet for the ADXL150/ADXL250
accelerometers, Rev. 0, 1998, available from Analog Devices, Inc.,
the contents of which is expressly incorporated by reference
herein.
[0050] Prior to discussing various pulse detection processes that
the defibrillator 10 may implement in accordance with the present
invention, a brief description of certain major components of the
defibrillator 10 is provided. Referring now to FIG. 5, the
defibrillator 10 includes defibrillation electrodes 30 (e.g.,
electrodes 12, 14 described above in FIG. 3). An accelerometer 26
(e.g., accelerometer 16 shown in FIG. 3) placed on the chest of the
patient produces electric signals in response to movement of the
chest wall. Depending on the output voltage of the accelerometer
26, a signal amplifier 28 may be provided to receive and amplify
the signal from the accelerometer 26 as appropriate for
digitization by analog-to-digital (A/D) converter 36. Prior to A/D
conversion, a filter 29 preferably filters the amplified
accelerometer signal to suppress noise and emphasize the portion of
the signal that most closely reveals chest wall movement due to
cardiac pulses in the patient.
[0051] The filtered accelerometer signal is delivered to the A/D
converter 36 which converts the signal into digital accelerometer
signal data for further evaluation. The filter 29 or other filters
(not shown) may also be provided to reduce any aliasing introduced
in the accelerometer signal by the A/D converter 36. The parameters
of such filtering depend, in part, on the sampling rate of the A/D
converter. Antialiasing filters, as well as A/D converters, are
well-known in the art, and may be implemented in hardware or
software, or a combination of both. For example, an embodiment of
the invention may use a hardware lowpass filter on the
accelerometer signal before the A/D converter 36, and then a
software highpass filter on the digital accelerometer signal data
after the A/D conversion. An additional software lowpass filter
after the A/D conversion may also be used to further limit the
bandwidth of the accelerometer signal data. In any respect, the A/D
converter 36 delivers the digital accelerometer signal data to a
processing circuit 38 for evaluation.
[0052] The processing circuit 38 evaluates the accelerometer signal
data for a feature indicating the presence of a cardiac pulse. The
processing circuit 38 is preferably comprised of a computer
processor that operates in accordance with programmed instructions
stored in a memory 40 that implement a pulse detection process 42,
described in more detail below. The processing circuit 38 may also
store in the memory 40 the accelerometer signal data obtained from
the patient, along with other event data and ECG signal data. The
memory 40 may be comprised of any type or combination of types of
storage medium, including, for example, a volatile memory such as a
dynamic random access memory (DRAM), a non-volatile static memory,
or computer-readable media such as a magnetic tape or disk or
optical storage unit (e.g., CD-RW or DVD) configured with permanent
or removable media.
[0053] The processing circuit 38 may report the results of the
pulse detection process to the operator of the defibrillator 10 via
a display 48. The processing circuit 38 may also prompt actions
(e.g., CPR) to the operator to direct the resuscitation effort. The
display 48 may include any kind of output device, for example,
lights, audible signals, alarm, printer, or display screen. The
processing circuit 38 may also receive input from the operator of
the defibrillator 10 via an input device 46. The input device 46
may include one or more keys, switches, buttons, dials, or other
types of user input devices.
[0054] The defibrillation electrodes 30 may further be used to
sense the patient's electrocardiogram (ECG) signals. ECG signals
obtained from the patient are amplified by the ECG signal amplifier
52 and filtered by the ECG bandpass filter 54 in a conventional
manner. The A/D converter 36 converts the ECG signals into
digitized ECG data and provides the ECG data to the processing
circuit 38 for evaluation.
[0055] Preferably, the processing circuit 38 evaluates the ECG
signals in accordance with programmed instructions 44 stored in the
memory 40 that carry out an ECG evaluation process to determine
whether a defibrillation shock should be provided. A suitable
method for determining whether to apply a defibrillation shock is
described in U.S. Pat. No. 4,610,254, which is assigned to the
assignee of the present invention and incorporated by reference
herein. If the processing circuit 38 determines that immediate
delivery of a defibrillation pulse is appropriate, the processing
circuit 38 instructs a defibrillation pulse generator 50 to prepare
to deliver the defibrillation pulse to the patient. In that regard,
the defibrillation pulse generator 50 uses an energy source (e.g.,
a battery) to charge one or more defibrillation capacitors in the
defibrillator 10.
[0056] When the defibrillation charge is ready for delivery, the
processing circuit 38 advises the operator via the display 48 that
the defibrillator 10 is ready to deliver the defibrillation pulse.
The processing circuit 38 may ask the operator to initiate the
delivery of the defibrillation pulse. When the operator initiates
delivery of the defibrillation pulse (e.g., via the input device
46), the processing circuit 38 instructs the defibrillation pulse
generator 50 to discharge through the patient the energy stored in
the defibrillation capacitors (via the defibrillation electrodes
30). Alternatively, the processing circuit 38 may cause the
defibrillation pulse generator 50 to automatically deliver the
defibrillation pulse when specified conditions (e.g., expiration of
a predetermined period of time, acceptable measured patient
impedance, etc.) are met.
[0057] In some circumstances, it may be preferable to apply CPR to
the patient before defibrillation even though cardiac conditions,
such as VF, are detected, especially for patients in whom
defibrillation is initially unlikely to succeed. See L. Cobb et
al., "Influence of Cardiopulmonary Resuscitation Prior to
Defibrillation in Patients with Out-of-Hospital Ventricular
Fibrillation" JAMA 281:1182-1188 (1999), incorporated by reference
herein. Thus, if desired, the defibrillator 10 may recommend the
application of chest compressions or CPR in situations where a
cardiac pulse is not detected and the ECG reveals a cardiac rhythm
for which immediate treatment by defibrillation therapy is not
indicated.
[0058] While FIG. 5 illustrates certain major components of the
defibrillator 10, those having ordinary skill in the art will
appreciate that the defibrillator 10 may contain more or fewer
components than those shown. The disclosure of a preferred
embodiment of the defibrillator 10 does not require that all of the
general conventional components be shown. It will further be
appreciated that aspects of the invention may be implemented in a
cardiac monitor having essentially the same components as the
defibrillator 10 shown in FIG. 5, except that the cardiac monitor
does not have the components necessary for delivering a
defibrillation pulse. Furthermore, some or all of the programmed
instructions 42 and 44 may be implemented in hardware as an
alternative to software instructions stored in the memory 40.
[0059] In any event, it is evident to one having ordinary skill in
the art that the present invention may be implemented by one or
more devices that include logic circuitry. The one or more devices
perform functions and/or methods as are described herein. The logic
circuitry may include a processor, such as the processing circuit
38, that may be programmable for a general purpose, or dedicated,
such as a microcontroller, a microprocessor, a digital signal
processor (DSP), etc. For example, a device implementing the
invention may be a digital computer-like device, such as a general
purpose computer selectively activated or reconfigured by a
computer program stored in the computer. Alternatively, the device
may be implemented as an application specific integrated circuit
(ASIC), etc.
[0060] The invention additionally provides methods and algorithms
that are described below. The methods and algorithms presented
herein are not necessarily inherently associated with any
particular computing device or other apparatus. Rather, various
general purpose machines may be used with programs in accordance
with the teachings herein, or it may prove more convenient to
construct more specialized apparatus to perform the required method
steps. The required structure for a variety of these machines
becomes apparent from this description.
[0061] In all cases, it should be borne in mind the distinction
between the method of the invention itself and the method of
operating a computing machine. The present invention relates to
both methods in general, and also to steps for operating a computer
and for processing electrical or other physical signals to generate
other desired physical signals.
[0062] The invention additionally provides programs and methods of
program operation. A program is generally defined as a group of
steps leading to a desired result. A program made according to an
embodiment of the invention is most advantageously implemented as a
program for a computing machine, such as a defibrillator 10 or
other equipment housing a general purpose computer, a special
purpose computer, a microprocessor, etc.
[0063] The invention also provides storage media that, individually
or in combination with others, have stored thereon instructions of
a program made according to the invention. A storage medium
according to the invention is a computer-readable medium, such as a
memory 40 as noted above, and is read by the computing machine
mentioned above.
[0064] It is readily apparent that the steps or instructions of a
program made according to an embodiment of the invention requires
physical manipulations of physical quantities. Usually, though not
necessarily, these quantities may be transferred, combined,
compared, and otherwise manipulated or processed according to the
instructions, and they may also be stored in a computer-readable
medium. These quantities include, for example, electrical,
magnetic, and electromagnetic signals, and also states of matter
that can be queried by such signals. It is convenient at times,
principally for reasons of common usage, to refer to these
quantities as signal data, bits, data bits, samples, values,
symbols, characters, images, terms, numbers, or the like. It should
be borne in mind, however, that all these and similar terms are
associated with the appropriate physical quantities, that these
terms are merely convenient labels applied to these physical
quantities.
[0065] This detailed description is presented largely in terms of
flowcharts, display images, algorithms, processes, and symbolic
representations of operations of data bits within at least one
computer readable medium. The present description achieves an
economy in that a single set of flowcharts is used to describe both
methods of the invention and programs according to the invention.
Such descriptions and representations are the type of convenient
labels used by those skilled in programming and/or data processing
arts to effectively convey the substance of their work to others
skilled in the art. A person skilled in the art of programming may
use these descriptions to readily generate specific instructions
for implementing a program according to the present invention.
[0066] Often, and for the sake of convenience only, it is preferred
to implement and describe a program as various interconnected
distinct software modules or features, individually and
collectively also known as software, though such modules may
equivalently be aggregated into a single program with unclear
boundaries. The software modules or features of the present
invention may be implemented by themselves, or in combination with
others. Although the program may be stored in a computer-readable
medium, such as a memory 40, a person skilled in the art will
readily recognize that it need not be a single memory, or even a
single machine. Various portions, modules, or features of the
program may reside in separate memories, or even separate machines.
The separate machines may be connected directly, or through a
network, such as a local area network (LAN), or a global network,
such as the Internet, by wired or wireless connections. For
example, a data acquisition unit may collect the accelerometer
signal data obtained in the present invention and communicate the
data to a remote computing machine for analysis and report whether
a cardiac pulse is present.
[0067] It will be appreciated that some of the methods described
herein may include software steps that can be performed by
different modules of an overall software architecture. For example,
data forwarding in a router may be performed in a data plane, which
consults a local routing table. Collection of performance data may
also be performed in a data plane. The performance data may be
processed in a control plane, which accordingly may update the
local routing table, in addition to neighboring ones. A person
skilled in the art will discern which step is performed in which
plane.
[0068] In any event, in the present case, methods of the invention
are implemented by machine operations. In other words, embodiments
of programs of the invention are made such that they perform
methods of the invention that are described in this document. These
may optionally be performed in conjunction with one or more human
operators performing some, but not all of them. As per the above,
these need not be co-located with each other, but each only with a
machine that houses a portion of the program. Alternatively, some
of these machines may operate automatically, without users and/or
independently from each other.
[0069] Methods of the invention are now described. In one aspect, a
pulse detection process conducted in accordance with the present
invention analyzes the accelerometer signal data obtained from the
patient to determine whether chest wall movement due to a cardiac
pulse is present in the patient. Characteristic vibrations of the
patient's chest are used as an indication of the presence of a
cardiac pulse in the patient. In another aspect, the pulse
detection process may analyze multiple physiological signals. For
example, the pulse detection process may analyze phonocardiogram
(PCG) data for heart sounds and impedance signal data for
characteristic fluctuations in patient impedance, combined with the
accelerometer signal data described herein, to determine the
presence of a cardiac pulse. See, e.g., the processing methods
described in the copending U.S. Patent Application titled PULSE
DETECTION APPARATUS, SOFTWARE, AND METHODS USING PATIENT
PHYSIOLOGICAL SIGNALS, filed concurrently herewith under Attorney
Docket No. PHYS118801, and incorporated by reference herein. A
combination of analyzed physiological signals may advantageously
provide a more robust pulse detection process with improved
detection characteristics.
[0070] FIG. 6 illustrates a pulse detection process 60a that
analyzes temporal energy in the accelerometer signal data. The
pulse detection process 60a begins at block 70 by obtaining signal
data from an accelerometer placed on a patient. As noted earlier,
signals received from an accelerometer placed on the patient (e.g.,
accelerometer 16 in FIG. 3) are converted into digital
accelerometer signal data.
[0071] The pulse detection process 60a evaluates the accelerometer
signal data for at least one feature indicative of the presence of
a cardiac pulse. In blocks 72 and 74, the pulse detection process
60a calculates estimates of the instantaneous energy and background
energy in the accelerometer signal data. The estimated
instantaneous energy may be calculated in block 72 simultaneously
with, before, or after, the calculation of estimated background
energy in block 74.
[0072] In block 72, the estimated instantaneous energy may be
calculated using a set of accelerometer signal data obtained from
the patient during a predetermined time window. One exemplary
embodiment of the invention uses a time window of 20 milliseconds
in length, though a longer, shorter, or shifted time window may be
used for estimating the instantaneous energy. The estimated
instantaneous energy may be calculated by squaring and summing each
of the accelerometer data values in the predetermined time
window.
[0073] The estimated background energy is calculated in block 74,
preferably using a set of accelerometer signal data obtained in an
earlier predetermined time window. One exemplary embodiment of the
invention calculates the estimated background energy using
accelerometer signal data in a 100 millisecond time window
commencing 220 milliseconds prior to the current time. The
accelerometer signal data within the earlier time window may also
be squared and summed to produce the estimated background energy.
Furthermore, other time window lengths and starting points may be
used.
[0074] The estimated instantaneous energy and background energy are
compared at block 76 to determine a relative change in energy in
the accelerometer signal data. The relative change in energy is
used by the pulse detection process 60a as a feature indicative of
the presence of characteristic chest vibrations, and hence the
presence of a cardiac pulse. If the relative change in energy
between the estimated instantaneous energy and the estimated
background energy exceeds a predetermined threshold, the pulse
detection process 60a determines that a cardiac pulse was present.
Because the calculation of background energy uses accelerometer
signal data obtained in a time window earlier than the
accelerometer signal data used to calculate instantaneous energy,
the rise and fall of the background energy waveform is expected to
generally follow the rise and fall of the instantaneous energy
waveform. Note that the background and instantaneous energies
should previously be normalized for purposes of comparison to each
other. For example, if squaring and summing is used and one energy
uses a 100 ms time window and the other energy uses a 20 ms time
window, the result of the energy using a 100 ms time window should
be divided by 5 so it can be properly compared against the result
from a 20 ms time window.
[0075] In decision block 78, if a cardiac pulse was detected, the
pulse detection process 60a proceeds to block 80 and reports the
presence of a cardiac pulse in the patient (thus indicating that
defibrillation therapy for the patient is not advised). Otherwise,
if a cardiac pulse was not detected, the pulse detection process
60a determines in block 82 that the patient is pulseless and that
defibrillation therapy may be appropriate. A defibrillator 10
implementing the pulse detection process 60a may proceed to
determine whether defibrillation therapy is appropriate, e.g., by
obtaining and processing ECG data from the patient as described in
U.S. Pat. No. 4,610,254, referenced earlier and incorporated herein
by reference.
[0076] In a further embodiment of the invention, the pulse
detection process 60a may be repeated over a specified time
interval or for a specified number of repetitions to produce a
series of determinations of whether a cardiac pulse is present in
the patient. The time windows for computing the estimated
instantaneous energy and background energy are shifted to
correspond with each instance of time in which the pulse detection
process 60a is performed. The pulse detection process 60a may
require a specified number of pulse detections before determining
that a cardiac pulse is in fact present in the patient.
[0077] During the time in which the instantaneous energy exceeds
the background energy by a predetermined threshold, the comparison
may return a "1", signifying the detection of a cardiac pulse. The
predetermined threshold may be adjusted to achieve a desired
sensitivity and specificity of detection. When the relative change
in energy between the instantaneous energy and the background
energy does not exceed the predetermined threshold, the comparison
may return a "0", signifying that a cardiac pulse has not been
detected.
[0078] FIG. 7 illustrates another pulse detection process 60b. As
with the detection process 60a, the detection process 60b analyzes
accelerometer signal data to detect the presence of characteristic
chest vibrations, and hence a cardiac pulse, in a patient. The
detection process 60b, however, focuses on a spectral energy
analysis of the accelerometer signal data (as compared to the
temporal energy analysis performed in the detection process
60a).
[0079] The pulse detection process 60b begins at block 100 by
obtaining accelerometer signal data from the patient in a manner as
discussed above with respect to block 70 (FIG. 6). In block 102,
the accelerometer signal data is preferably analyzed to identify a
set of accelerometer signal data that likely contains information
identifying the presence of a cardiac pulse. In that regard, the
candidate accelerometer data may be identified by using the
temporal energy comparison discussed in block 76 of the pulse
detection process 60a. When the estimated instantaneous energy
exceeds the estimated background energy by a predetermined
threshold, the energy comparison suggests that a cardiac pulse has
been detected. Alternatively, a set of accelerometer signal data
potentially identifying a cardiac pulse may be selected by
evaluating the patient's ECG data for the occurrence of an R-wave.
The timing of cardiac pulse vibrations in the patient's chest in
relation to an R-wave is generally known in the art and may be used
to predict the timing of candidate data in the accelerometer signal
data. Other embodiments of the invention may compute an energy
spectrum without first identifying candidate accelerometer data,
e.g., by continuously computing an energy spectrum using the most
current accelerometer data as the candidate data.
[0080] Next, in block 104, the pulse detection process 60b computes
an energy spectrum of the candidate accelerometer signal data,
preferably using a maximum entropy method, though other spectral
calculations may be used. Computing an energy spectrum using a
maximum entropy method ("MEM spectrum") is well-known in the art.
See, e.g., Modern Spectral Estimation: Theory and Application, by
Stephen M. Kay, published by Prentice Hall of Englewood Cliffs,
N.J., beginning at p. 182, and incorporated herein by reference. An
MEM spectrum typically appears smoother than an energy spectrum
produced by Fourier transform techniques. The MEM spectrum may be
normalized by removing a baseline (e.g., DC) energy value across
the MEM spectrum.
[0081] The frequency of a peak energy value in the energy spectrum
may be used as a feature indicative of the presence of a cardiac
pulse. The frequency of the selected peak is evaluated against a
predetermined threshold frequency value to decide whether a cardiac
pulse has been detected. In block 106 (FIG. 7), the pulse detection
process 60b evaluates the energy values in the MEM spectrum to
identify a peak value in the MEM spectrum and determine its
frequency.
[0082] In block 108, the frequency of the peak value is compared
with a predetermined threshold frequency to decide whether a
cardiac pulse is detected. For example, if the frequency of the
peak is less than or equal to a threshold frequency, e.g., 100 Hz,
the pulse detection process 60b determines that a cardiac pulse was
detected. Alternative embodiments of the invention may use values
other than 100 Hz for the predetermined threshold frequency.
[0083] If a cardiac pulse was detected, the pulse detection process
60b proceeds from decision block 110 to block 112 and determines
that a pulse is present in the patient, thus advising against
application of a defibrillation pulse. If, in decision block 110, a
cardiac pulse was not detected, the pulse detection process 60b
determines in block 114 that the patient is pulseless and that
defibrillation may be appropriate for the patient. In that case,
further signal processing of ECG data obtained from the patient is
preferably performed to determine the applicability of
defibrillation therapy, e.g., as described in U.S. Pat. No.
4,610,254, referenced earlier. In some circumstances, CPR therapy
is warranted.
[0084] FIG. 8 illustrates another pulse detection process 60c that
also uses an MEM spectrum as calculated in block 104 of the
detection process 60b. Instead of analyzing the frequency of a peak
value in the MEM spectrum, as performed in the process 60b, the
process 60c analyzes the energy of a peak value in the MEM
spectrum.
[0085] The detection process 60c begins at block 150 by obtaining
accelerometer signal data from the patient in a manner as discussed
earlier with respect to block 70 (FIG. 6). The accelerometer signal
data is analyzed in block 152 to identify candidate accelerometer
signal data corresponding to the time when a cardiac pulse likely
occurred. The analysis performed in block 152 may include an energy
comparison process or ECG analysis as described earlier with
respect to block 102 of pulse detection process 60b (FIG. 7). An
MEM spectrum of the candidate accelerometer signal data is then
computed in block 154 in a manner as discussed earlier with respect
to block 104 (FIG. 7). Also as noted before, the energy spectrum
calculation process may run continuously.
[0086] In block 156, the pulse detection process 60c evaluates the
energy values in the MEM spectrum to locate a peak value in the
spectrum. The energy value of the peak, determined in a block 158,
is used as a feature indicative of the presence of a cardiac pulse,
and is compared in block 160 with a predetermined threshold energy
to decide whether a cardiac pulse was detected. If the energy of
the peak value exceeds the threshold energy, the pulse detection
process 60c determines in decision block 162 that a cardiac pulse
was detected.
[0087] If, in decision block 162, a cardiac pulse was detected, the
pulse detection process 60c proceeds to block 164 and determines
that a cardiac pulse is present in the patient. In that
circumstance, the detection process 60c may advise against
providing defibrillation therapy to the patient. The detection
process may also advise to check patient breathing. On the other
hand, if a cardiac pulse was not detected in decision block 162,
the pulse detection process 60c proceeds to block 166 and
determines that the patient is pulseless. In that circumstance, the
detection process 60c advises that defibrillation therapy may be
appropriate for the patient. In other embodiments, a prompt that
advises the application of chest compressions or CPR may be given
in addition to or in place of advising defibrillation therapy for
pulseless patients. An analysis of ECG data, as noted earlier, may
be used to determine the applicability of defibrillation
therapy.
[0088] On occasion, it is possible that noise in the accelerometer
signal data may cause a false detection of what appears to be
characteristic chest vibrations, and hence false detection of a
cardiac pulse, when using one of the detection processes 60
described herein. If the signal-to-noise ratio of the accelerometer
signal data obtained from the patient is not high enough to avoid
such false detection of a cardiac pulse, the pulse detection
processes 60 may be combined in one or more ways to produce a pulse
detection process with improved specificity. For example, FIG. 9
illustrates a detection process 60d that combines aspects of the
detection processes 60a, 60b, and 60c.
[0089] In FIG. 9, the pulse detection process 60d begins at block
170 by obtaining accelerometer signal data from a patient, e.g., in
a manner as described earlier with respect to block 70 of pulse
detection process 60a (FIG. 6). After the accelerometer signal data
is obtained, estimates of the instantaneous energy and the
background energy in the accelerometer signal data are computed in
blocks 172 and 174, e.g., in a manner as described earlier with
respect to blocks 72 and 74. The estimated instantaneous and
background energy values are then compared in a block 176, e.g., as
described earlier with respect to block 76, to produce a first
detection statistic, or feature, indicative of the presence of a
cardiac pulse. The first detection statistic produced in block 176
is provided to a multidimensional classifier in block 186 that
evaluates detection statistics to determine whether a cardiac pulse
has been detected. Alternatively, the instantaneous and background
energies computed in blocks 172 and 174 may be directly provided as
separate detection statistics to the multidimensional classifier in
block 186 for joint classification with any other detection
statistics provided to the classifier (i.e., eliminating the
comparison performed in block 176).
[0090] The accelerometer signal data obtained in block 170 is also
used in identifying candidate data that is likely indicative of a
cardiac pulse and for computing an MEM spectrum of the candidate
data in block 178, in a manner as described earlier with respect to
blocks 102 and 104 of pulse detection process 60b (FIG. 7). Once
the MEM spectrum is computed, the pulse detection process 60d in
block 180 locates a peak value in the MEM spectrum.
[0091] The frequency of the peak value is determined in a block 182
and provided as a second detection statistic, or feature, to the
classifier in block 186. Alternatively, the second detection
statistic may be the result of comparing the frequency of the peak
value with a threshold frequency, e.g., in a manner as described
earlier with respect to block 108 (FIG. 7), to produce the second
detection statistic.
[0092] In block 184, the pulse detection process 60d also
determines the energy at the peak value and provides the energy
value as a third detection statistic, or feature, to the classifier
in block 186. The peak energy value may alternatively be compared
with a threshold energy, e.g., in a manner as described earlier
with respect to block 160 (FIG. 8), to produce the third detection
statistic.
[0093] The classifier in block 186 jointly classifies the first,
second, and third detection statistics using a multidimensional
classifier to determine whether a cardiac pulse is present in the
patient. Techniques for constructing multidimensional classifiers
are well-known in the art. For an expanded description of
classifiers suitable for use in the present invention, see, e.g.,
R. Duda and P. Hart, Pattern Classification and Scene Analysis,
published by John Wiley & Sons, New York, and incorporated
herein by reference.
[0094] The classifier in block 186 may also use a voting scheme to
determine whether a cardiac pulse is present in the patient. For
example, if any of the first, second, or third detection statistics
indicates the detection of a cardiac pulse (e.g., the instantaneous
energy exceeded the background energy by a threshold value, the
frequency of a peak was equal to or less than a threshold
frequency, or the energy of the second peak exceeded a threshold
energy), the classifier may determine that a pulse is present in
the patient. Alternatively, the classifier in block 186 may
determine that a pulse is present by finding that a combination of
the first, second, and third detection statistics is indicative of
the presence of a cardiac pulse (e.g., a positive indication from
the first detection statistic combined with a positive indication
from the second or third detection statistics, etc.). The
classifier in block 186 may also weight the first, second, or third
detection statistics to emphasize one detection statistic over
another in deciding whether a cardiac pulse is present.
[0095] If, in decision block 188, a cardiac pulse was detected, the
pulse detection process 60d determines in block 190 that a pulse is
present in the patient and may advise the operator of the
defibrillator to not defibrillate the patient. The process may also
advise to not perform CPR, in connection with or in place of any
defibrillation advice. Otherwise, if a cardiac pulse was not
detected in decision block 188, the pulse detection process 60d
determines in block 192 that the patient is pulseless and that
CPR/chest compressions and/or defibrillation therapy may be
appropriate. An analysis of ECG data, as described earlier in
reference to U.S. Pat. No. 4,610,254, may be used to determine
whether defibrillation therapy is appropriate.
[0096] An analysis of ECG data may also be combined with an
analysis of accelerometer signal data to determine the presence of
a cardiac pulse in the patient. In one aspect, detecting a
ventricular complex, such as a QRS complex, in the ECG data in time
relation to the occurrence of a characteristic feature in the
accelerometer signal data may serve to confirm the detection of a
cardiac pulse. In another aspect, detecting a ventricular complex
in the ECG data may be used to identify accelerometer signal data
for use in the pulse detection process, since a characteristic peak
in the accelerometer signal data is expected to occur in time
proximity to the occurrence of a ventricular complex if a cardiac
pulse is present in the patient. This aspect of the invention is
also helpful in identifying whether the patient is in a state of
pulseless electrical activity. If a ventricular complex is found in
the ECG data and a characteristic peak or other feature indicating
a cardiac pulse does not occur in the accelerometer signal data
within an expected time period, the patient may be considered in a
state of pulseless electrical activity (PEA) which may be reported
to the operator of the device. The operator may also be prompted to
deliver PEA-specific therapy to the patient.
[0097] FIG. 10 illustrates another pulse detection process 60e that
analyzes accelerometer signal data obtained during time intervals
associated with ventricular complexes (e.g., QRS complexes) in the
patient's ECG. Beginning in block 202, the pulse detection process
60e captures both ECG and accelerometer signal data, synchronized
in time, for a predetermined time interval (e.g., 10 seconds).
Alternatively, the ECG and accelerometer signal capturing step may
continue until the first or a specified number of QRS complexes in
the ECG have been identified, or in the event of asystole or a low
heart rate, a predetermined maximum period of time (e.g., 10
seconds) has passed. During this time, persons around the patient
should be advised to not touch the patient (e.g., the device could
report "Analyzing now . . . Stand clear").
[0098] In block 204, the pulse detection process 60e locates QRS
complexes in the ECG signal. Identification of QRS complexes can be
done using methods published in the literature and well-known to
those skilled in the art of ECG signal processing. For example see,
Watanabe K., et al., "Computer Analysis of the Exercise ECG: A
Review," Prog Cardiovasc Dis 22: 423-446, 1980.
[0099] In block 206, for each time that a QRS complex was
identified in the ECG signal, a segment of accelerometer signal
data obtained from the patient is selected. In one embodiment of
the invention, the time window of each segment of accelerometer
signal data is approximately 600 milliseconds in length, and
commences in time slightly before the QRS complex. If no QRS
complexes were identified in the captured ECG signal in block 204
(as would happen for example, during asystole), no segments of
accelerometer signal data are selected in block 206.
[0100] In block 208, one or more measurements are made on a segment
of accelerometer signal data selected in block 204 to identify or
calculate a feature indicative of a cardiac pulse. Non-limiting
examples of the measurements may include one or more of the
following temporal parameters:
[0101] (1) peak-to-peak amplitude of the accelerometer signal data
in the segment;
[0102] (2) peak-peak amplitude of a derivative of the accelerometer
signal data in the segment;
[0103] (3) energy of the accelerometer signal in the segment
(preferably calculated by squaring and summing each of the data
values in the segment); or
[0104] (4) a pattern matching statistic.
[0105] The previously-described instantaneous/background energy
methods, as well as the spectral methods described herein, may be
used in block 208 as well to identify or calculate a feature
indicative of a cardiac pulse. As to pattern matching, the segment
of accelerometer signal data is compared with one or more
previously identified accelerometer signal patterns known to
predict the presence of a pulse. The comparison produces a pattern
match statistic. Generally, in this context, the greater the value
of the pattern match statistic, the closer the patient's
accelerometer signal matches a pattern accelerometer signal that
predicts the presence of a pulse. A measurement resulting from the
analysis in block 208 constitutes a feature of the accelerometer
signal data that may be indicative of the presence of a pulse.
[0106] In decision block 210, the one or more features from block
208 are evaluated to determine the presence of a cardiac pulse in
the patient. The process 60e shown in FIG. 10 compares the one or
more features to predetermined thresholds to determine whether or
not a pulse is detected. For example, a peak-to-peak amplitude
measurement would be consistent with the presence of a pulse if the
measurement exceeded a predetermined threshold. Similarly, an
energy measurement would be consistent with a pulse if its
magnitude exceeded a predetermined threshold. Likewise, a pattern
matching statistic would be consistent with a pulse if it exceeded
a predetermined threshold. If the feature exceeded the specified
threshold, the pulse detection process 60e determines that a pulse
was detected, as indicated at block 212. If the feature did not
exceed the specified threshold, a pulse was not detected, as
indicated at block 214. If no segments of accelerometer signal data
were selected in block 206 (i.e., no QRS complexes were located in
block 202 in the captured ECG), the pulse detection process 60e
would determine that a pulse was not detected, as indicated at
block 214.
[0107] While thresholding is used in block 210 to determine whether
a pulse was detected, those skilled in the art will recognize other
forms of classification that may suitably be used in the invention.
For example, a multidimensional classifier may be used in decision
block 210 to determine whether a pulse was detected. Separate
analyses of the amplitude and energy in the accelerometer data
segment may be performed, with the resultant outcome of each
analysis constituting a detection statistic that is provided to the
multidimensional classifier. The detection statistics may be
weighted and compared in the classifier to determine an overall
conclusion whether a pulse is present in the patient. In other
embodiments, individual calculations of instantaneous and
background amplitudes and/or energies may be provided as detection
features for evaluation in a multidimensional classifier. Pattern
match statistics may also be evaluated in the multidimensional
classifier, as may other measurements of the accelerometer signal
data. Furthermore, spectral techniques can be used, such as the
peak frequency or energy techniques previously described.
Techniques for constructing multidimensional classifiers are known
in the art. See, e.g., R. Duda and P. Hart, Pattern Classification
and Scene Analysis, referenced earlier and incorporated herein by
reference.
[0108] After determining whether a pulse was detected (block 212)
or not detected (block 214), the pulse detection process 60e
determines whether all of the segments of accelerometer signal data
selected in block 206 have been analyzed. If not, the analysis and
decision process of blocks 208, 210, 212, and 214 is preferably
repeated for a new accelerometer data segment. This continues until
all of the accelerometer data segments selected in block 206 have
been analyzed.
[0109] The resulting determination (pulse detected or no pulse
detected) may not be the same for each accelerometer data segment
analyzed. An additional decision step is used to determine the
overall outcome of the pulse detection process 60e. As indicated at
decision block 218, the pulse detection process 60e may evaluate
the determinations for each accelerometer signal data segment and
decide that a pulse is present in the patient if a pulse was
detected in a simple majority of the segments analyzed. Of course,
other voting schemes may be used. If, in decision block 218, a
majority is found, the pulse detection process concludes that a
cardiac pulse is present in the patient, as indicated at block 220.
Otherwise, the pulse detection process 60e concludes that the
patient is pulseless, as indicated at block 222.
[0110] Requiring a pulse to be found in more than a simple majority
of the accelerometer data segments would improve the specificity of
the detection, but decrease the sensitivity for detecting a pulse.
Conversely, requiring a pulse to be found for just one
accelerometer data segment or for less than a majority of the
accelerometer segments would improve sensitivity for detecting a
pulse but decrease specificity. If the pulse detection process 60e
concludes that a pulse is present in the patient, the process 60e
may optionally proceed to check the pulse rate of the patient, as
illustrated in FIG. 11.
[0111] Turning to FIG. 11, in block 224, the number of QRS
complexes (located in block 204 in FIG. 10) are counted. Decision
block 226 subsequently compares the number of QRS complexes to a
threshold. In one exemplary embodiment, the threshold is 5,
corresponding to a heart rate of approximately 30 bpm. If the
number of QRS complexes is at least equal to the threshold, the
pulse detection process 60e proceeds to block 228, concluding that
the patient has a pulse and an adequate pulse rate. If the number
of QRS complexes is less than the threshold, the pulse detection
process 60e proceeds to block 230, concluding that the patient has
a pulse, but also severe bradycardia. At very low heart rates,
however, the blood flow may be insufficient to support life. For
that reason, below a certain heart rate (e.g., 30 bpm), the patient
may instead be considered pulseless.
[0112] While the pulse detection process shown in FIG. 10 includes
capturing both ECG and accelerometer signal data, and selecting
segments of accelerometer signal data based on ventricular
complexes located in the ECG, other pulse detection processes may
not capture or use the ECG signal. In FIG. 12, an alternative pulse
detection process 60f begins by capturing only accelerometer signal
data from the patient, as indicated at block 234. Depending on the
length of the time interval in which accelerometer signal data is
captured, it may be advantageous to select a segment of the
accelerometer signal data for further analysis, as indicated at
block 236. In that regard, one suitable selection process includes
scanning the accelerometer signal data for a peak value and
selecting a segment of data that surrounds the detected peak.
[0113] For exemplary purposes, the pulse detection process 60f is
shown evaluating the selected segment of accelerometer signal data
using a pattern match analysis. However, those skilled in the art
will recognize that other techniques (e.g., analysis of the
amplitude or energy--temporal or spectral--in the accelerometer
signal data, as discussed above) may be used. In block 238, the
selected accelerometer data segment is compared with previously
identified accelerometer signal patterns known to predict the
presence of a pulse. The resulting pattern match statistic is
evaluated against a threshold in decision block 240 to determine
whether a pulse was detected in the patient. If the pattern match
statistic exceeded the threshold, the pulse detection process 232
concludes in block 241 that a pulse was detected in the patient.
Otherwise, the pulse detection process 232 concludes that the
patient is pulseless, as indicated in block 242. At this point, the
pulse detection process is finished. Alternatively, if a pulse was
detected in the patient, the pulse detection process 232 may
proceed to evaluate the patient's pulse rate in a manner described
in reference to FIG. 11.
[0114] The accelerometer signal obtained from the sensor placed on
the patient may include signal elements that are due to cardiac
pulse vibrations, respiration, or other patient motion. To assess
whether a patient has a pulse, it is desirable to suppress elements
in the accelerometer signal that are due to causes other than
cardiac pulses. Signal elements due to noncardiac causes may
contain components at frequencies similar to those due to cardiac
pulses. Consequently, bandpass filtering may not always adequately
suppress accelerometer signals due to noncardiac causes.
[0115] Signal averaging of the accelerometer signal can be used to
suppress signal elements that are due to noncardiac causes. Signal
averaging makes advantageous use of the fact that accelerometer
signal elements due to cardiac pulse vibrations are generally
synchronized to ventricular complexes in the ECG signal, whereas
other signal elements are generally asynchronous to ventricular
complexes. Pulse detection may be more accurately accomplished
using an averaged accelerometer signal.
[0116] One preferred method for averaging the accelerometer signal
first stores the continuous ECG and accelerometer signals,
synchronized in time, for a predetermined time interval (e.g., 10
seconds). The timing of the QRS complexes (if any) in the stored
ECG signal are determined. Using true mathematical correlation (or
an alternative correlation technique such as area of difference),
the QRS complexes are classified into types, where all QRS
complexes of the same type have high correlation with the first
occurring QRS complex of that type. The dominant QRS type is
selected as the type containing the most members, with a preference
for the narrowest QRS type when a two or more types tie for most
members. Using the first QRS of the dominant type as a reference
complex, the second QRS complex of the same type is shifted in time
until it is best aligned with the reference complex (i.e., it
achieves a maximum correlation value). The corresponding
accelerometer signal is also shifted in time to stay synchronized
with the time-shifted QRS complex. When the second QRS complex is
optimally aligned with the reference complex, the two QRS complexes
are averaged together. Segments of the corresponding accelerometer
signals, over a time period from slightly before the start of the
QRS complex to about 600 milliseconds after the end of the QRS
complex, are also averaged together. The averaged QRS complex is
then used as a new reference complex and the process of averaging
both the QRS complexes and the corresponding accelerometer data is
repeated with the remaining QRS complexes of the dominant type.
[0117] Preferably, during the subsequent averaging of the QRS
complexes and accelerometer data segments, the new QRS complex and
accelerometer segment carry a weight of one and the previous
averaged QRS complex and accelerometer segment carry a weight equal
to the number of QRS complexes that have been included in the
averaged QRS complex. When all of the QRS complexes of the dominant
type have been processed as described above, the averaged
accelerometer signal segment is evaluated using one or more of the
techniques previously described (e.g., amplitude, energy, pattern
matching) to determine whether the patient has a pulse.
[0118] Averaging of accelerometer data segments may also be
accomplished without ECG data. For example, segments of
accelerometer data may be analyzed and classified into types where
segments of the same type have a high correlation. Accelerometer
data of a dominant type, for example, may then be averaged,
evaluated as previously described (using amplitude, energy, pattern
matching, etc.) to determine whether the patient has a pulse.
[0119] During severe bradycardia, there will be few QRS complexes
in a 10-second period and signal averaging of the accelerometer
signal will not be as effective as when the heart rate is higher.
However, at very low heart rates, there is unlikely to be enough
blood flow to support life. For that reason, below a certain heart
rate (e.g., 30 bpm), the patient may be considered pulseless.
[0120] A pulse detection process as described herein may be used as
part of an overall shock advisory process in a defibrillator. The
shock advisory process determines whether to recommend
defibrillation or other forms of therapy for a patient. FIG. 13
illustrates a pulse detection/defibrillation process 260,
preferably for use in an automated external defibrillator (AED)
capable of providing a defibrillation pulse if a patient is
determined to be pulseless and in ventricular fibrillation or
ventricular tachycardia.
[0121] In the pulse detection/defibrillation process 260, an AED
initializes its circuits when it is first turned on, as indicated
at block 262. The defibrillation electrodes of the AED are placed
on the patient. When the AED is ready for operation, the process
260 performs an analysis of the patient, as indicated at block 264,
in which the AED obtains selected information such as accelerometer
signal data and/or ECG data from the patient. During the analysis
performed in block 264, the AED preferably reports "Analyzing now .
. . Stand clear" to the operator of the AED.
[0122] Using the information obtained in the patient analysis, the
process 260 determines in decision block 266 whether the patient is
experiencing ventricular fibrillation (VF). If VF is present in the
patient, the process 260 proceeds to block 276 where the AED
prepares to deliver a defibrillation pulse to the patient. In that
regard, an energy storage device within the AED, such as a
capacitor, is charged. At the same time, the AED reports "Shock
advised" to the operator of the AED.
[0123] Once the energy storage device is charged, the process 260
proceeds to block 278 where the AED is ready to deliver the
defibrillation pulse. The operator of the AED is advised "Stand
clear . . . Push to shock." When the operator of the AED initiates
delivery of the defibrillation pulse, the process 260 delivers the
defibrillation shock to the patient, as indicated in block 280.
[0124] The AED preferably records in memory that it delivered a
defibrillation pulse to the patient. If the present pulse delivery
is the first or second defibrillation shock delivered to the
patient, the process 260 may return to block 264 where the patient
undergoes another analysis. On the other hand, if the pulse
delivery was the third defibrillation pulse to be delivered to the
patient, the process 260 may proceed to block 274 where the AED
advises the operator to commence providing CPR therapy to the
patient, e.g., by using the message "Start CPR." The "No shock
advised" prompt shown in block 274 is suppressed in this instance.
The AED may continue to prompt for CPR for a predetermined time
period, after which the patient may again be analyzed, as indicated
in block 264.
[0125] Returning to decision block 266, if VF is not detected in
the patient, the process 260 proceeds to decision block 268 and
determines whether a cardiac pulse is present in the patient. The
pulse detection performed in block 268 may be any one or a
combination or variation of the pulse detection processes described
above.
[0126] Breathing may be checked manually by the operator or
automatically by the device, as discussed below in regard to block
374 of FIG. 15. If, at decision block 268, a pulse is detected in
the patient and the patient is not breathing, the process 260
proceeds to block 270 and reports "Pulse detected . . . Start
rescue breathing" to the operator. The process 260 may also report
"Return of spontaneous circulation" if a pulse is detected in the
patient any time after the delivery of a defibrillation pulse in
block 280. In any event, after a predetermined time period for
rescue breathing has completed, the process 260 preferably returns
to block 264 to repeat an analysis of the patient.
[0127] If a cardiac pulse is not detected at decision block 268,
the process 260 determines whether the patient is experiencing
ventricular tachycardia (VT) with a heart rate of greater than a
certain threshold, e.g., 100 beats per minute (bpm), as indicated
at decision block 272. Other thresholds such as 120, 150, or 180
bpm, for example, may be used. If the determination at decision
block 272 is negative, the process 260 proceeds to block 274 and
advises the operator to provide CPR therapy. Again, at this point,
the AED reports "No shock advised . . . Start CPR" to the operator.
The prompt to provide CPR is preferably provided for a defined
period of time. When the period of time for CPR is finished, the
process 260 preferably returns to block 264 and performs another
analysis of the patient. If the determination at decision block 272
is positive (i.e., the patient is experiencing VT with a heart rate
greater than the threshold), the process 260 performs the shock
sequence shown at blocks 276, 278, 280 to deliver a defibrillation
pulse.
[0128] Those having ordinary skill in defibrillation and cardiac
therapy will recognize variations and additions to the process 260
within the scope of the invention. FIG. 14, for example,
illustrates an alternative pulse detection/defibrillation process
300 for use in an AED. As with the process 260 in FIG. 15, the AED
begins by initializing its circuits at block 302. At block 304, the
AED performs an analysis of the patient in a manner similar to that
described with respect to block 264 in FIG. 13. After completing
the analysis of the patient, the process 300 proceeds to decision
block 306 to determine whether a pulse is present in the patient.
The pulse detection performed in block 306 may be, for example, any
one of the pulse detection processes discussed above or a
combination or variation thereof.
[0129] If a pulse is detected in the patient, the process 300 may
enter a monitoring mode at block 308 in which the patient's pulse
is monitored. The pulse monitoring performed at block 308 may use
any one or a combination of the pulse detection processes described
above. Preferably, the process 300 is configured to proceed from
block 308 to block 304 after expiration of the predetermined
monitoring time period. If the pulse monitoring at block 308
determines at any time that a pulse is no longer detected, the
process 300 returns to block 304 to perform another analysis of the
patient. The process 300 also preferably reports the change in
patient condition to the operator.
[0130] If, at decision block 306, a pulse is not detected in the
patient, the process 300 proceeds to decision block 310 where it
determines whether the patient has a shockable cardiac rhythm
(e.g., VF or VT). As referenced earlier, U.S. Pat. No. 4,610,254,
incorporated herein by reference, describes a suitable method for
differentiating shockable from non-shockable cardiac rhythms.
[0131] If a shockable cardiac rhythm, such as VF or VT, is
detected, the process 300 proceeds to a shock delivery sequence at
blocks 312, 314, and 316, which may operate in a manner similar to
that described with respect to blocks 276, 278, and 280 in FIG. 13.
If the pulse delivery was the third defibrillation shock delivered
to the patient, the process 300 may proceed to block 318 and prompt
the delivery of CPR, as discussed with block 274 in FIG. 13.
[0132] If VF or VT is not detected at decision block 310, the
process 300 checks for asystole, as indicated at block 320. One
suitable process for detecting asystole is described in U.S. Pat.
No. 6,304,773, assigned to the assignee of the present invention
and incorporated herein by reference. If asystole is detected at
block 320, the process 300 proceeds to prompt the delivery of CPR,
as indicated at block 318. If asystole is not detected, the process
300 determines that the patient is experiencing pulseless
electrical activity (PEA), as indicated at block 322. PEA is
generally defined by the presence of ventricular complexes in a
patient and the lack of a detectable pulse, combined with no
detection of VT or VF. Detection of PEA in block 322 is achieved by
ruling out the presence of a pulse (block 306), detecting no VF or
VT (block 310), and detecting no asystole (block 320).
Alternatively, if the ECG signal is monitored for ventricular
complexes (e.g., as shown at block 202 in FIG. 10), the process 300
may conclude the patient is in a state of PEA if it repeatedly
observes ventricular complexes without detection of a cardiac pulse
associated therewith. If a PEA condition is detected, the process
300 proceeds to block 324 and prompts the operator to deliver
PEA-specific therapy to the patient. One suitable method of
treating PEA is described in U.S. Pat. No. 6,298,267, incorporated
by reference herein. The process 300 may prompt other therapies as
well, provided they are designed for a PEA condition. After a
PEA-specific therapy has been delivered to the patient, possibly
for a predetermined period of time, the process 300 returns to
block 304 to repeat the analysis of the patient.
[0133] FIG. 15 illustrates yet another pulse
detection/defibrillation process 350 that may be used in an AED. At
block 352, after the AED has been turned on, the AED initializes
its circuits. The defibrillation electrodes are also placed on the
patient. The AED is then ready to analyze the patient, as indicated
at block 354. This analysis may be performed in a manner similar to
that described with respect to block 264 in FIG. 13.
[0134] If at any point the AED determines that the defibrillation
electrodes are not connected to the AED, the process 350 jumps to
block 356 where the AED instructs the operator to "Connect
electrodes." When the AED senses that the electrodes are connected,
the process 350 returns to the analysis in block 354. Likewise, if
the AED finds itself in any other state where the electrodes are
not connected, as represented by block 358, the process 350 jumps
to block 356 where it instructs the operator to connect the
electrodes.
[0135] Furthermore, during the analysis performed in block 354, if
the AED detects motion on the part of the patient, the process 350
proceeds to block 360 where the AED reports to the operator of the
AED "Motion detected . . . Stop motion." If the patient is moved
during the analysis process 354, the data obtained during the
analysis is more likely to be affected by noise and other signal
contaminants. Motion of the patient may be detected in an
impedance-sensing signal communicated through the patient. A
suitable method for detecting motion of the patient is described in
U.S. Pat. No. 4,610,254. The AED evaluates the impedance measured
between the defibrillation electrodes placed on the patient. Noise
and signal components resulting from patient motion cause
fluctuations in the impedance signal, generally in a frequency
range of 1-3 Hz. If the measured impedance fluctuates outside of a
predetermined range, the AED determines that the patient is moving
or being moved and directs the process 350 to proceed to block 360.
When the motion ceases, the process 350 returns to the analysis in
block 354.
[0136] The process 350 next proceeds to decision block 362 where it
determines whether a pulse is detected in the patient. Again, the
pulse detection processes performed in decision block 362 may be,
for example, one of the pulse detection processes described above
or combination or variation thereof.
[0137] If a pulse is not detected in the patient, the process 350
proceeds to decision block 364 where it determines whether the
patient has a shockable cardiac rhythm (e.g., VF or VT) or a
non-shockable cardiac rhythm (such as asystole and bradycardia). As
referenced earlier, one suitable method for differentiating
shockable from non-shockable cardiac rhythms is disclosed in U.S.
Pat. No. 4,610,254. If the patient's cardiac rhythm is determined
to be shockable (e.g., VF or VT is found), the process 350 proceeds
to blocks 366, 368, and 370 to deliver a shock to the patient. The
shock delivery may be performed as described earlier with respect
to blocks 276, 278, 280 in FIG. 13.
[0138] If the pulse delivery was the third defibrillation pulse to
be delivered to the patient, the process 350 proceeds to block 372
where the AED advises the operator to commence providing CPR
therapy to the patient. The CPR prompt may continue for a defined
period of time, at which the process 350 returns to block 354 and
performs another analysis of the patient.
[0139] If, at decision block 364, the patient's cardiac rhythm is
determined not shockable, the process 350 preferably proceeds to
block 372 and advises the operator to provide CPR therapy, as
discussed above.
[0140] Returning to decision block 362, if a pulse is detected in
the patient, the process 350 proceeds to decision block 374 where
it determines whether the patient is breathing. In that regard, the
AED may use the impedance signal for determining whether a patient
is breathing. Fluctuations in patient impedance below 1 Hz are
largely indicative of a change in volume of the patient's lungs.
The breathing detection at block 374 (and at blocks 376 and 378,
discussed below) may monitor the impedance signal for
characteristic changes that indicate patient breathing, e.g., as
described in Hoffmans et al., "Respiratory Monitoring With a New
Impedance Plethysmograph," Anesthesia 41: 1139-42, 1986, which is
incorporated by reference herein. Detection of breathing may employ
a process that evaluates an amplitude, energy, or pattern in the
impedance signal. Preferably, a bandpass filter would be used to
isolate the frequency components that more closely demonstrate
patient breathing. The accelerometer data may also be analyzed for
a component that reveals whether the patient's body is moving due
to breathing. If automatic means for detecting breathing in the
patient are not available, the AED may ask the operator of the AED
to input information (e.g., by pressing a button) to indicate
whether the patient is breathing.
[0141] If, at decision block 374, the process 350 determines that
the patient is not breathing, the process 350 proceeds to a block
376 where the operator of the AED is advised to commence rescue
breathing. In that regard, the AED reports to the operator "Pulse
detected . . . Start rescue breathing." The AED also continues to
monitor the patient's cardiac pulse and returns to block 354 if a
cardiac pulse is no longer detected. If, at any point during the
provision of rescue breathing, the AED detects that the patient is
breathing on his own, the process 350 proceeds to block 378 where
the AED monitors the patient for a continued presence of breathing
and a cardiac pulse.
[0142] Returning to decision block 374, if the process 350
determines that the patient is breathing, the process 350 proceeds
to block 378 where the AED monitors the pulse and breathing of the
patient. In that regard, the AED reports "Pulse and breathing
detected. . . Monitoring patient." If, at any time during the
monitoring of the patient the process 350 determines that the
patient is not breathing, the process 350 proceeds to block 376
where the operator of the AED is advised to commence rescue
breathing. If a cardiac pulse is no longer detected in the patient,
the process 350 proceeds from either block 376 or 378 to block 354
to commence a new analysis of the patient.
[0143] Lastly, as noted in FIG. 15, during the rescue breathing
procedure in block 376 or the monitoring procedure performed in
block 378, the AED may assess whether CPR is being administered to
the patient. In that regard, signals received from the
accelerometer 16 shown in FIG. 3 may be used to measure parameters,
such as frequency and depth of chest compressions being applied to
the patient. If the AED finds that CPR is being performed, the AED
may prompt the operator to cease providing CPR. If, during the CPR
period of block 372, the AED determines that CPR is not being
administered to the patient, the AED may remind the operator to
provide CPR therapy to the patient. Another method for determining
whether CPR is being administered is to monitor patient impedance
to observe patterns of impedance fluctuation in the patient that
are indicative of CPR. During CPR, repetitive chest compression
typically causes repetitive fluctuations in the impedance
signal.
[0144] FIG. 16 illustrates yet another application in which pulse
detection according to the present invention may be used. The
application described in FIG. 16 pertains to auto-capture detection
in cardiac pacing.
[0145] Specifically, the auto-capture detection process 380 begins
at block 382 in which pacing therapy for the patient is initiated.
A counter N, described below, is set to equal 0. At block 384, a
pacing pulse is delivered to the patient. Thereafter, accelerometer
signal data is obtained from the patient, as indicated at block
386. The accelerometer signal data is used in block 388 to detect
the presence of a cardiac pulse. The pulse detection process used
in block 388 may be, for example, any one or combination or
variation of the pulse detection processes discussed above.
[0146] The sequence of delivering a pacing pulse and determining
the presence of a cardiac pulse in blocks 384, 386, 388 may be
repeated a number of times. With respect to FIG. 16, for example,
the sequence is repeated five times. At block 390, the counter N is
evaluated, and if not yet equal to 5, the counter is incremented by
1 (block 392), following which the process 380 returns to deliver
another pacing pulse to the patient (block 384).
[0147] If, at decision block 390, the counter N equals 5, the
process 380 determines at decision block 394 whether a cardiac
pulse occurred consistently after each pacing pulse. The process
380 requires that some portion or all of the pacing pulses result
in a detectable cardiac pulse before pronouncing that capture has
been achieved. If the presence of a cardiac pulse is determined to
consistently follow the pacing pulses, the process 380 determines
that capture has been achieved, as in indicated at block 396.
Otherwise, the current of the pacing pulses is increased by a
predetermined amount, e.g., 10 milliamperes, as indicated at block
398. At block 399, the counter N is set back to equal 0 and the
process 380 returns to the pacing capture detection sequence
beginning at block 384. In this manner, the pacing current is
increased until capture has been achieved.
[0148] In FIG. 16, the presence of a pulse is used to determine
whether the pacing stimulus has been captured by the ventricles of
the patient's heart. Detection of ventricular complexes in the
patient's ECG may also be used in connection with accelerometer
signal data to identify pacing capture. For example, a ventricular
complex will occur immediately following the pacing stimulus if
capture has been achieved. If ventricular complexes are not
observed, the current of the pacing pulses may be increased, as
discussed above, until capture has been achieved. In an alternative
embodiment, a user of the device may be prompted to increase the
current of the pacing stimuli prior to the pacing stimuli current
being increased.
[0149] FIG. 17 illustrates still another application in which pulse
detection according to the present invention may be used. The
process 400 described in FIG. 17 is particularly suited for use in
a manual defibrillator or patient monitor, though it may be
implemented in other forms of medical devices. Beginning at block
402, the process 400 monitors the patient's ECG for QRS complexes.
At block 404, the process 400 also obtains accelerometer signal
data from the patient. The process 400 uses the ECG and
accelerometer signal data in decision block 406 to determine the
presence of a cardiac pulse. The pulse detection implemented in
block 406 may be one or a combination or variation of the pulse
detection processes discussed herein.
[0150] If a pulse is detected, the process 400 determines whether a
defibrillation pulse has been provided to the patient and if so,
reports the return of spontaneous circulation to the operator, as
indicated at block 418. The process 400 then returns to block 402
to repeat the pulse detection analysis. If a pulse is not detected,
the process 400 evaluates the ECG signal to determine whether the
patient is experiencing ventricular fibrillation or ventricular
tachycardia with a heart rate greater than 100 bpm. If so, then the
process identifies the patient's condition and produces a VT/VF
alarm, as indicated at block 410. If not, the process 400 then
proceeds to block 412 to check for an asystole condition.
[0151] Detection of asystole may be accomplished as noted earlier
and described in U.S. Pat. No. 6,304,773, incorporated herein by
reference. If asystole is detected, the process 400 identifies the
patient's condition and sounds an asystole alarm, as indicated at
block 414. Otherwise, the patient is experiencing PEA and the
patient's condition is so identified, with the sound of a PEA
alarm, as indicated at block 416. In this manner, the operator of
the manual defibrillator or monitor is kept advised of the
patient's condition.
[0152] While various exemplary embodiments of the invention have
been illustrated and described herein, persons having ordinary
skill in the art will recognize variations of the same that are
fully with the scope of the invention. Embodiments of the invention
described herein are shown processing digital accelerometer signal
data. However, the invention also includes embodiments in which the
accelerometer signal data is not converted to digital form, but
remains in analog form. References to "data" thus encompass both
digital and analog signal formats. Moreover, references to
"accelerometer signal data" may refer to the raw accelerometer
signal itself or signal information derived from the accelerometer
signal in either digital or analog form.
* * * * *